Release Notes for Cisco IOS Release 12.1(7)DA3 for Cisco DSLAMs with NI-2

These release notes describe features and caveats in Cisco IOSRelease 12.1(7)DA3 for the Cisco 6015, Cisco 6100, Cisco 6130, Cisco 6160, and Cisco 6260 digital subscriber line access multiplexers (DSLAMs).

Note When you upgrade from Cisco IOS 12.1(5)DA1 or earlier images on the NI-2 card to Release 12.1(7)DA3, you must format the bootflash on the NI-2 card before loading the 12.1(7)DA3 dboot image. See the "Update the Image in Bootflash to the 12.1(7)DA3 dboot Image" section.

For pointers to more information about the Cisco 6015, Cisco 6100, Cisco 6130, Cisco 6160, Cisco 6260, and their software, refer to the "Related Documentation" section. To learn more about caveats, visit the Cisco web site—See the "Obtaining Documentation" section for details. Information about electronic documentation can be found in both the "Obtaining Documentation" section and the "Documentation CD-ROM" section.

This document contains the following topics:

System Requirements

Cisco IOS Release 12.1(7)DA3 runs on these DSLAM systems:

Cisco 6015
Cisco 6160
Cisco 6260

New and Changed Information

This section provides new and changed information for Cisco IOS releases.

New Features in Release 12.1(7)DA3

No new features were introduced in Cisco IOS Release 12.1(7)DA3.

New Features in Release 12.1(7)DA

This section describes new features in Cisco IOS Release 12.1(7)DA.

E1 I/O Module on the Cisco 6260

This release adds support for E1 I/O module with eight individual E1 interfaces that can be allocated as either individual E1 trunk or subtend interfaces, or as inverse multiplexing over ATM (IMA) group trunk or subtend interfaces. Also, the E1 I/O module provides the alarm pinouts for the Cisco 6260 system. See the T1/E1 Inverse Multiplexing over ATM feature module for further information.

New Quad-Port DMT ATU-C Line Card Support

This release adds support for the quad-port DMT ATU-C line card (4xDMT) to control a received power attenuator. The dsl atuc-1-4dmt rx-attenuation configuration turns on and off a received power attenuator in all the 4xDMTs in a chassis. Setting this configuration also automatically retrains all the ports on all the 4xDMTs.

You can determine the state of this setting by using the show running configuration command to check for no dsl atuc-1-4dmt rx-attenuation. The default setting, dsl atuc-1-4dmt rx-attenuation, is not displayed by show running configuration.

NI-2 Card and APS Link Redundancy

Cisco IOS Release 12.1(7)DA adds support for NI-2 card redundancy and Synchronous Optical Network Automatic Protection Switching (SONET APS) on the Cisco 6130, Cisco 6160, and Cisco 6260 DSLAM systems. The redundancy feature has two main components.

Recovery from failure on an NI-2 card lets the standby card take over if the active card fails.
Recovery from failure due to a cut fiber or the failure of an optical transmitter or receiver (SONET APS) lets the active NI-2 switch to the fiber interface on the standby NI-2 card.

For detailed information, see the NI-2 Card and APS Link Redundancy feature module.

Octal-Port G.SHDSL SHTU-C Line Card

The octal-port G.SHDSL SHTU-C line card (8xG.SHDSL) is designed for use in the Cisco 6015, Cisco 6160, and Cisco 6260 without a POTS splitter configuration. An 8xG.SHDSL port is physically connected to one specific subscriber line through the main distribution frame (MDF) connections.

The 8xG.SHDSL:

Supports eight G.SHDSL modem connections.
Supports trellis coded pulse amplitude modulation (TC-PAM) line encoding.
Converts G.SHDSL modulation from the line into digital data streams to and from the NI-2 card.
Determines the bit rate for customer premises equipment (CPE).

Installation Notes

Update the Image in Bootflash to the 12.1(7)DA3 dboot Image

When you upgrade from Cisco IOS Release 12.1(5)DA1 or earlier images on the NI-2 card to Release 12.1(7)DA3, you must reformat the bootflash on the NI-2 card.

Note If you are upgrading from Cisco IOS Release 12.1(6)DA, 12.1(6)DA2, or 12.1(7)DA you do not need to reformat the bootflash. You may skip this section and follow the DSLAM upgrade procedure in http://www.cisco.com/univercd/cc/td/doc/product/dsl_prod/ios_dsl/rel121/config/load.htm.

Note Cisco highly recommends that you have console access to the NI-2 card during the upgrade procedure. You can use the console connection to troubleshoot any unexpected events that occur during the upgrade.

To update the boot image, perform the following steps in privileged EXEC mode:

	Command	Purpose
Step 1	`DSLAM# dir bootflash:`	Verify that the bootflash image is ni2-dboot-mz.121-5.da1 or ni2-dboot-mz.121-4.da. If it is neither, go to the "Upgrade the Bootflash Image on the NI-2 Card to the 12.1(5)DA1 dboot Image" section and perform the instructions there, then return to thisstep.
Step 2	`DSLAM#dir flash:`	Display the name of the flash file that begins ni2- and use it as filename in Step 3.
Step 3	`DSLAM#delete flash:filename`	Delete the flash file name found in Step 2. Repeat Step 2 and Step 3 until all files in the flash have beendeleted.
Step 4	`DSLAM#squeeze flash:`	Recover available space in Flash memory.
Step 5	`DSLAM#copy tftp://tftpserver:TFTPBOOT/ni2-dsl-mz.1 21-7.da3 flash:`	Copy the Cisco IOS image from a tftp server toFlash.
Step 6	`DSLAM#configure terminal DSLAM (config)#no boot system DSLAM (config)#boot system flash:ni2-dsl-mz.121-7.da3 DSLAM#end`	Enter global configuration mode. Disable the boot from system. Specify the name of the system image to load at start-up. End global configuration mode.
Step 7	`DSLAM#copy running-config startup-config`	Save your changes to the start-up configuration.
Step 8	`DSLAM#reload`	Reload the system to upgrade the image.
Step 9	`DSLAM#show version`	Confirm that the running image is ni2-dsl-mz.121-7.da3. If it is not, go to Step 5.
Step 10	`DSLAM# format bootflash:`	Erase all information in bootflash memory. Answer y to all confirm questions. When the DSLAM returns you to the EXEC prompt, bootflash memory is successfully formatted and ready for use. Ensure that the bootflash is 3.8 MB total. If it is not, go to Step 5.
Step 11	`DSLAM#copy tftp://tftpserver:TFTPBOOT/ni2-dboot-mz .121-7.da3 bootflash:`	Copy the boot image from a tftp server to thebootflash.
Step 12	`DSLAM# reload`	Reload the system to upgrade the image.

Upgrade the Bootflash Image on the NI-2 Card to the 12.1(5)DA1 dboot Image

When you upgrade from Release 12.1(3)DA or earlier images on the NI-2 card to Release 12.1(5)DA1, Cisco recommends that you upgrade the bootflash image on the NI-2 card to the 12.1(5)DA1 dboot image. This is a result of a problem (CSCdr89374) in old bootflash images. This defect may cause the flash application image to fail to self-decompress successfully.

To upgrade the dboot image in bootflash, perform the following steps:

	Command	Purpose
Step 1	`DSLAM>enable Password: <password> DSLAM#`	Enter enable mode. Enter the password. You have entered enable mode when the prompt changes to DSLAM#.
Step 2	`DSLAM# delete bootflash:filename`	Make room in the bootflash by deleting the name of the current boot image.
Step 3	`DSLAM# squeeze bootflash`	Recover available space in bootflash using the squeeze bootflash command.
Step 4	`DSLAM# copy tftp://[server name] /[directory]/ni2-dboot-mz.121-5.da1 bootflash:`	Copy the boot image to the bootflash.
Step 5	`DSLAM# show version`	Record the current value of the config-register that appears on the last line of the show version display.
Step 6	`DSLAM# configure terminal`	Enter global configuration mode. You have entered global configuration mode when the prompt changes to `DSLAM (config)#`.
Step 7	`DSLAM(config)# config-register 0`	Set the to 0x0000 so that the NI-2 will reboot in the rom monitor.
Step 8	`DSLAM(config)# exit`	Exit configure terminal mode.
Step 9	`DSLAM# copy running-config startup-config`	Save the running configuration.
Step 10	`DSLAM# reload`	Reset the system.
Step 11	`rommon> set`	If you see BOOTLDR after you enter this command, the image in bootflash is already being used as the bootstrapper; go to Step 15. Otherwise, go to Step 12 and enter the commands in Steps 12 through 14 to force the system to use ni2-dboot-mz.121-5.bin as the bootstrapper.
Step 12	`rommon > unset BOOTLDR`	Unset BOOTLDR to remove the variable.
Step 13	`rommon > sync`	Sync to save the state of rommon.
Step 14	`rommon > b`	When the NI-2 boots, ituses ni2-dboot-mz.121-5da.bin as yourbootstrapper.
Step 15	`DSLAM# configure terminal`	Enter global configuration mode. You have entered global configuration mode when the prompt changes to `DSLAM (config)#`.
Step 16	`DSLAM(config)# config-register value`	Set the config-register to the value you recorded in Step5.
Step 17	`DSLAM(config)# exit`	Return to enable mode. Go to the "Update the Image in Bootflash to the 12.1(7)DA3 dboot Image" section.

Limitations and Restrictions

This section describes the limitations and restrictions for Cisco IOS DSLAM releases.

Cisco IOS Release 12.1(7)DA

This section describes the limitations and restrictions for Cisco IOS Release 12.1(7)DA.

Frame Relay PVCs/Soft PVCs on an IDSL Interface

In Cisco IOS Release 12.1(7)DA, the number of Frame Relay PVCs/Soft PVCs on an IDSL interface is restricted to one if you use the default row in frame-relay connection traffic table (FR-CTT).

When upgrading to 12.1(7)DA from previous releases, you must first create a new row in the FR-CTT with the desired CIR value and use the resultant row number during PVCs/Soft PVCs creation. If you do not create a new row, the second FR PVCs/Soft PVCs command is not parsed and installed on the IDSL interface.

Cisco IOS Release 12.1(6)DA

This section describes the limitations and restrictions for Cisco IOS Release 12.1(6)DA. Other restrictions that apply to Release 12.1(6)DA are described in the "Cisco IOS Release 12.1(5)DA" section.

Use of a Ring Topology in a DSLAM Subtend Environment to Achieve ATM Trunk Redundancy

Ring topology is achieved when a node in the subtend tree is attached to the ATM access to provide a physical redundant loop. Thus, when the primary ATM access or one of the ATM trunks in the subtend tree fails, the soft permanent virtual circuits (SPVCs) can be dynamically rerouted through the use of private network-network interface (PNNI). See Figure 1.

Figure 1: SPVCs Prior To Failure

This redundancy requires the use of SPVCs. If you use permanent virtual circuits (PVCs) or permanent virtual paths (PVPs), redundancy cannot be provided. The use of the SPVCs allows traffic to be rerouted around the failed access point, because SPVCs leverage this feature of PNNI. When the failure occurs, the SPVCs are disconnected and dynamically reconnected across the new path. See Figure 2.

Figure 2: SPVCs Dynamically Rerouted

Requirements

If you implement ring topology in a DSLAM subtend to achieve ATM trunk redundancy, the following requirements apply:

You must use SPVCs, PNNI, ATM signaling, and Interim Local Management Interface (ILMI) to enable rerouting. PNNI, ATM signaling, and ILMI are enabled by default. Permanent connections such as PVCs and shaped virtual paths (VPs) does not benefit from the redundant link.
You must make the redundant link's PNNI administrative weight higher than the PNNI weight of the primary trunk. Once you change the weight of the redundant link, the subtend tree uses this link only if a failure occurs.

Limitations

When the redundant link is active, the following occurrences are problems:

Loss of subtending fairness.
Increase in latency as well as an increase in cell delay variation (CDV) between the cells. Delay-sensitive traffic, such as voice and video, or traffic that is susceptible to jitter, such as constant bit rate (CBR) voice, might be comprised. This technique is best realized for unspecified bit rate (UBR) traffic, such as consumer internet access, where no strict Quality of Service (QoS) objectives are required.

Once the redundant link is active, the following occurrences are potential problems:

Greater possibility of increased congestion in the DSLAM ATM switch fabric, which may cause loss of data.
When the main link is restored, there will be downtime while the path is being rerouted. Once the SPVCs are rerouted to the redundant ATM trunk and the original trunk is repaired or brought back into service, you must manually intervene. You must flap (shut/noshut) the subtend port. Because of retries on the current path, you must keep the trunk down until the maximum retry interval expires.

Cisco IOS Release 12.1(5)DA

There are no restrictions that specifically apply just to Cisco IOS Release 12.1(5)DA. Other restrictions that apply to Cisco IOS Release 12.1(5)DA are described in the "Cisco IOS Release 12.1(4)DA" section.

Cisco IOS Release 12.1(4)DA

This section describes the limitations and restrictions for Cisco IOS Release 12.1(4)DA.

Restrictions to MPLS VPN Mapping of RFC 1483 Routed Sessions

This section describes restrictions to Cisco MPLS VPN mapping of RFC 1483 routed sessions.

Note DSLAMs running MPLS VPN mapping of RFC 1483 sessions are referred to as IP DSL switches.

Number of Configurable MPLS VPNs Limited to 50

Each DSLAM can support up to 50 MPLS VPNs.

Integrated Routing and Bridging Not Supported

MPLS VPN mapping of RFC 1483 routed sessions must not be confused with Integrated Routing and Bridging (IRB). IRB is not supported by MPLS VPN mapping of RFC 1483 routed sessions.

VPN Interfaces Restricted to Trunk Interfaces

Do not configure subtended interfaces for MPLS VPN services. Only trunk interfaces support MPLS VPN mapping of RFC 1483 routed sessions.

MPLS ATM-Label Switch Router Functionality Not Supported

DSLAMs are not meant for use as MPLS ATM-Label Switch Routers (ATM-LSRs). When designing your network, keep in mind that DSLAMs act only as Label Edge Routers (LER).

Performance Restrictions for MPLS VPN Traffic

MPLS VPN-enabled interfaces do not perform as well as switched VCs. Please take this into consideration when deploying MPLS VPNs in your networks.

Restricted Layer 3 Services

The following Layer 3 services are not supported in Cisco IOS Release 12.1(4)DA:

IP Quality of Service
IP Queueing
IP Multicast

Restricted MPLS Features

The following are MPLS-related features that are not a part of the MPLS VPN mapping of RFC 1483 routed sessions feature:

MPLS Traffic Engineering
MPLS Multicast

DSL Interface Limitations

In IP DSL switches, each DSL interface can support multiple Permanent Virtual Circuits (PVCs), but only one routed MPLS VC.

Configuration of MPLS VPN Mapping of RFC 1483 Routed Sessions not Supported by Cisco DSL Manager (CDM)

CDM users can provision switched VCs, but CDM does not support configuring routed termination of RFC 1483 sessions.

MPLS VPN Mapping of RFC 1483 Routed Sessions not Supported on the Eight-Port IDSL ITU-C Line Card

Routed termination of IDSL connections is not supported in Cisco IOS Release 12.1(4)DA.

Cisco IOS Release 12.1(3)DA

The limitations and restrictions for Cisco IOS Release 12.1(3)DA are the same as those listed for Cisco IOS Release 12.1(2)DA.

Cisco IOS Release 12.1(2)DA

The following restrictions apply to Release 12.1(2)DA:

1. If you are using Cisco DSL Manager (CDM) Release 1.0 or 1.1 software to manage your existing DSLAM network:

a. Do not upgrade the DSLAMs to this new Cisco IOS Release. CDM Release 1.0 and CDM Release 1.1 do not support this release of the Cisco IOS software.

b. Spare NI-2 cards are shipped with the most recent release of Cisco IOS software. Therefore, if you install a spare NI-2 card in a network that uses CDM, you might need to install an older release of the Cisco IOS software on the spare NI-2 card. Cisco IOS Release 12.0(8)DA4 and earlier releases are compatible with CDM. Go to the Software Center on CCO to download back releases of Cisco IOS software.

c. When CDM Release 3.0 comes out, it will support the 4xflexsi and 4xSDSLs only. CDM 3.0 Release will not allow you to configure or manage 8xIDSLs. Configuration and management of 8xIDSLs will be supported in a later release of CDM.

d. Refer to your CDM documentation for more information on compatibility between CDM releases and Cisco IOS releases.

2. Cisco IOS Release 12.0(8)DA2 or greater is required for Cisco 6260 and Cisco 6130 systems that contain new 4xDMTs (revision 800-05262-03 or greater). Releases 12.0(8)DA2 and above are also compatible with earlier revisions of the 4xDMT.

If you need to fall back to an earlier release, note the following restrictions:

a. You can fall back only to Release 12.0(8)DA1 or greater. The DSLAMs cannot operate under releases prior to 12.0(8)DA1.

b. New 4xDMTs (revision 800-05262-03 or greater) require Cisco IOS Release 12.0(8)DA2 or greater; the cards will not operate under prior releases of software.

3. Cisco IOS Release 12.0(8)DA4 runs only on this hardware:

a. Cisco 6260

b. Cisco 6130 with NI-2 card and Cisco 6130-specific system I/O cards

4. The VP tunneling feature cannot be used in conjunction with VCs outside of VP tunnels.

Cisco IOS Release 12.1(1)DA1

The limitations and restrictions for Cisco IOS Release 12.1(1)DA1 are the same as those listed for Cisco IOS Release 12.1(1)DA, below.

Cisco IOS Release 12.1(1)DA

The following restrictions apply to this release:

1. If you are using Cisco DSL Manager (CDM) software to manage your existing DSLAM network:

a. Do not upgrade the DSLAMs to Release 12.1(1)DA. No CDM release is available to work with this release of the DSLAM software.

If you need to fall back to an earlier release, note the following restrictions:

a. You can fall back only to Release 12.0(8)DA1 or greater. The DSLAMs cannot operate under releases prior to 12.0(8)DA1.

b. New 4xDMTs (revision 800-05262-03 or greater) require Cisco IOS Release 12.0(8)DA2 or greater; the cards will not operate under prior releases of software.

3. Cisco IOS Release 12.0(8)DA4 runs only on this hardware:

a. Cisco 6260

b. Cisco 6130 with NI-2 card and Cisco 6130-specific system I/O cards

4. The VP tunneling feature cannot be used in conjunction with VCs outside of VP tunnels.

Important Notes

This section provides important information about Cisco IOS DSLAM releases.

IDSL Linecard Clocking Problem Fixed in Release 12.1(4)DA

The following problem is fixed in the Cisco IOS 12.1(4)DA release. Upgrade to 12.1(4)DA to fix this IDSL problem. The 8xIDSL is not forwarding traffic after reload/reboot because the ISDN chipset is initializing indefinitely due to network clock synchronization.

Soft PVC Address Changes Upon Upgrade to 12.1(3)DA or 12.1(4)DA

When you upgrade from a previous release to 12.1(3)DA or 12.1(4)DA, the default soft PVC addresses on all interfaces change. This only occurs when you upgrade to release 12.1(3)DA or 12.1(4)DA from a previous release. This will not happen when you upgrade to future releases from 12.1(4)DA.

Workarounds:

Reconfigure the soft PVCs associated with all interfaces.
Assign a (nondefault) address to the interfaces.

Configuring Cisco Routers for Use with IDSL

If you wish to use a Cisco router for an IDSL application and the router is running a Cisco IOS release prior to 12.1, you must configure the ISDN switch type. If you do not configure the ISDN switch type on the Cisco router, the router's BRI interface may not come back up after the IDSL line goes down and back up.

To prevent this problem from occurring, execute the isdn switch-type basic-ni command in global configuration mode on the router.

This problem does not occur if the Cisco router is running Cisco IOS Release 12.1 or later.

Upgrading from Cisco IOS 12.0(x)DA to 12.1(1)DA, 12.1(2)DA, or 12.1(3)DA

After you upgrade a DSLAM from Cisco IOS Release 12.0(x)DA to 12.1(1)DA, 12.1(2)DA, or 12.1(3)DA, you might experience a problem that results from a change in some of the DMT defaults in the DSL profiles. The changes affect the training mode, overhead framing, and DMT margin parameters. This problem is fixed in Release 12.1(4)DA.

The problem is triggered when you edit your DSL configuration profiles after upgrading to Cisco IOS Release 12.1(x)DA.

The system checks its startup configuration file for a release number, and if it finds a number earlier than 12.1, it automatically adds the following commands to the profile whenever you edit the profile. These commands set the profile to use the old default values for training mode, overhead framing, and DMT margin.

dmt training-mode standard
dmt overhead-framing mode1
dmt margin downstream 3 upstream 3

The commands are added so that the changed defaults in Cisco IOS Release 12.1(x)DA will not affect your configuration. However, the system continues to add the commands to your configuration every time you edit profiles, unless your startup configuration file contains a 12.1 release stamp.

Workaround: Follow these steps. Start with Step 2 if you have already completed the upgrade.

Step 1 Reboot the DSLAM after you upgrade to Cisco IOS Release 12.1(x)DA or higher.

Step 2 Issue the command copy running-config startup-config. Now the startup configuration file has a 12.1(x)DA stamp in it.

Step 3 Reboot the DSLAM again.

Now you can change the configuration without difficulty. [CSCdr66665]

Assigning VPI Values to Shaped VP Tunnels

This release supports the full range of VPI values: 0 to 255. However, if you configure VP tunnels with traffic shaping, you can use only 32 VPIs out of that range. If you have not yet assigned any VPIs, all values from 0 to 255 are available. Once you start assigning VPIs, however, the assigned VPIs limit the VPIs that remain. (You assign VPIs using the atm pvp or atm pvc commands.)

After a particular VPI value is assigned to a shaped VP tunnel, every 32nd VPI value above and below the first one is eliminated—that is, the original value modulo 32. For example, if you assign VPI 94 to a shaped VP tunnel, the following VPI values become unavailable for any purpose: 30, 62, 126, 158, 190, and 222.

To avoid problems, choose a block of 32 consecutive VPI values (for example, 0 to 31 or 101 to 132). The software rejects invalid VPI values.

Installing Multiple Cisco 6160 Chassis in an Equipment Rack

You can install multiple Cisco 6160 chassis within a telco equipment rack. A standard seven-foot equipment rack can house four Cisco 6160 chassis, stacked one on top of another. This configuration maximizes the DSL density within a seven-foot rack. However, if space is available or if you are interested in using multiservice capabilities that the chassis will support in the future, Cisco recommends that you install no more than three Cisco 6160 chassis within a seven-foot rack. Leave a space of at least 2.5 rack units (4.375 inches, or 11.1 cm) beneath each chassis for future cable management use.

Synchronizing the Management Database

If you use both the Cisco DSL Manager (CDM) application and the command line interface (CLI) to configure and manage your DSLAMs, you must the synchronize the CDM to your management database on a regular basis. (If you use only the CDM or only the CLI, you need not synchronize your database.) Refer to the release notes for the Cisco DSL Manager 2.0 or later for more information.

Console Logging

Turn console logging off if you plan to reboot the DSLAM. Turn console logging back on after the system comes up. (Console logging is turned on by default.) Use the global configuration commands no logging console (to turn the feature off) and logging console (to turn it on).

If console logging is on when the system reboots, the large volume of console messages consumes CPU time. As a result, the system comes back up more slowly and line cards might reload repeatedly, causing further delays.

Caveats

This section describes unexpected behavior in the system. A tracking number, if one exists, is provided in square brackets at the end of each description. Please refer to the tracking number if you communicate with Cisco Systems about any of these issues.

Open Caveats—Release 12.1(7)DA2

MIB Caveats

Unexpected Reload When Querying cIdslPppoa tables

DSLAM performs an unexpected reload whenever an SNMP getNext request is issued for the cIdslPppoaPerfDataTable, cIdslPppoaIntervalTable or the cIdslPppoaPerfDataTable. (CISCO-IDSL-LINE-MIB) Applies only to IDSL interfaces under the following conditions:

SNMP getNext of the above tables when the interfaces are attached to a profile that has IDSL encapsulation set to frame-relay.

Workaround: None. [CSCdu60701]

NI-2 Card Redundancy Caveats

Autosync of Running Config Can Cause Write Command to Fail

In a redundant configuration, if running-config sync is enabled, the startup config can be lost when a write command is issued immediately after a configure terminal session ends. The reason is that the write command tries to access the running configuration while it is locked because it is being sent to the standby card.

Workaround: Disable automatic running-config sync and use the manual sync facility. [CSCdu35414]

Open Caveats—Release 12.1(7)DA

The caveats in this section apply only to Cisco IOS Release 12.1(7)DA. Other open caveats that apply to this Release 12.1(7)DA are described in the "Open CaveatsRelease 12.1(6)DA" section, the "Open CaveatsRelease 12.1(5)DA" section, the "Open CaveatsRelease 12.1(4)DA" section, the "Open CaveatsRelease 12.1(3)DA" section, the "Open CaveatsRelease 12.1(2)DA" section, "Open CaveatsRelease 12.1(1)DA1" section, and the "Open CaveatsRelease 12.1(1)DA" section.

4xDMT over ISDN Caveats

Some DSL Statistics Missing in 12.1(5)DA to 12.1(7)DA for DMT Cards

Some DSL Statistics are missing on the show dsl interface atm command display: LOS Events, LOF/RFI Events ES Events: NOT Present for DMT.

Workaround: None. [CSCdt88209]

DMT Profile Caveats

Minimum DMT Bit Rate Setting Lost After Upgrade

The minimum bit rate settings configured for a DMT profile are lost after an upgrade. This happens only when the configured minimum bit rate setting values are more than 640 and 128 for downstream and upstream respectively.

Workaround: Reconfigure the minimum bit rate settings for the respective DMT profile. [CSCdt71353]

DSL Profile Caveats

Profile Created from Altered Default has Diff Parameters

Created profiles do not inherit modified default profile parameters.

Workaround: Create and change profile parameters for desired interface operation. Do not rely on modified default profile inheritance. [CSCdt34090]

IMA Caveats

E1 IMA Group Continuously Loses Cells When Physical Link Removed

Interoperability with the Cisco LS1010 can cause cell loss on Cisco 6015/Cisco 6260 if an E1 IMA group consists of either 1 or 3 active links.

Workaround: None. [CSCdt49728]

4xFlexi-DMT Line Card Caveats

Upstream Errors on Some Alcatel CPE in Fast Mode

Alcatel-based CPEs show upstream errors when upstream is configured above 864 kbps in the fastpath.

Workaround: Set the maximum upstream bit rate to 864 kbps or less. [CSCdt59614]

8xG.SHDSL Caveats

In Very Rare Cases after Power Cycle a Port Will not Pass Data

In very rare cases after a power cycle of the chassis, a port will not pass data upstream.

Workaround: To get the port out of this condition, perform a shut/no shut on the port. If using SNMP, set ifAdminStatus for the port to down(2) and the set the ifAdminStatus for the port to up(1). [CSCdr79925]

Far End Statistics Reported as Zero

For the 8xG.SHDSL line card, all far end (STU-R) statistics displayed through CLI or through SNMP are reported as zero.

Workaround: None. [CSCds51529]

In Extremely Rare Cases After a Power Cycle, Port does not Train

In extremely rare cases after a power cycle, a port will not train.

Workaround: To get the port to train, perform a shut/no shut on the port. If using SNMP to manage the system, set the ifAdminStatus to down(2) and then set the ifAdminStatus to up(1). [CSCds52657]

Attenuation Reporting Zero Through CLI and SNMP

When displaying the attenuation for a trained 8xG.SHDSL line via CLI or with the MIB, the value for the attenuation is reported as zero.

Workaround: None. [CSCds70283]

1544 Annex B Line Rate Does not Meet PSD Mask Limit

At the line rate of 1544 Mbps Annex B, the 8xG.SHDSL line card exceeds the symmetric PSD Mask and transmit power as specified in G.992.1 specification.

Workaround: There is no workaround for this DDTS when the rate is set to 1544 Annex B. [CSCdt72363]

The 8xG.SHDSL Line Card does not Support Rate Adaption

The 8xG.SHDSL line card does not support rate adaption.

Workaround: None. [CSCdt87954]

Active LED on 8xDMT & 8xG.SHDSL Stays on When NI-2 Communication Lost

When the 8xDMT or 8xG.SHDSL establishes initial communication with the NI-2, the active LED on the line card face plate will be illuminated. If communication with the NI-2 is lost subsequent to this, through a reload command or removal of the NI-2 card, the active LED will remain illuminated.

Workaround: None. [CSCdu12337]

8xIDSL Caveats

IDSL Traffic Stops when a Second NI-2 Card is Inserted in a Chassis

When a second NI-2 card is inserted into a live chassis, 8xIDSLs briefly lose sync with the network clock, causing them to drop active IDSL connections. Network clock sync is automatically restored when the second NI boots into standby mode and IDSL connections come back up automatically.

Workaround: None. [CSCds25811]

Link Down Traps not Reliably Sent for IDSL

Link down trap is not sent consistently for IDSL interfaces.

Workaround: None. [CSCdt75675]

MIB Caveats

No Log Event Link-status has no Effect

LINK and LINEPROTO event log message appears on the console whenever the interface state changes either from UP to DOWN or vice versa, even when logging is disabled at the interface.

Workaround: None. [CSCds14884]

Link up Traps Still Mapped to ATM Interface

Link up traps are mapped to ATM layer.

Workaround: Use ifStackTable to determine physical layer. [CSCdt55085]

The Delay does not Change when Changing oamLoopbackPingDelay.

The delay does not change when setting oamLoopbackPingDelay.

Workaround: None. [CSCdt73955]

adslAturChanTransmittedBlks has a MIB value of 0 while CLI has non-0

Getting adslAturChanTransmittedBlks returns 0.

Workaround: Use value from CLI. [CSCdt79352]

adslAturChanReceivedBlks has a MIB value of 0 and CLI has non-0

Getting adslAturChanReceivedBlks returns 0.

Workaround: Use value from CLI. [CSCdt79381]

adslAturVendorID will return 0 from MIB, CLI has 57 (Globespan)

Getting adslAturVendorID returns 0.

Workaround: Use value from CLI. [CSCdt79387]

adslAtucChanCrcBlockLength always has a value of 0

Getting adslAtucChanCrcBlockLength always returns 0.

Workaround: None. [CSCdt79394]

adslAturChanCrcBlockLength always has a value of 0

Getting adslAturChanCrcBlockLength always returns 0.

Workaround: None. [CSCdt79405]

cIdslItucInvSerialNumber returns garbage instead of empty string

Getting cIdslItucInvSerialNumbe returns garbage instead of blank string.

Workaround: None. [CSCdt79412]

NI-2 Card Redundancy Caveats

On OC-3 NI-2s, PPP Sessions May Drop During Switchovers

NI-2 switchovers interrupt data traffic for approximately 60 seconds. If CPE devices or aggregation points are configured to time out PPP sessions in a similar time frame, the sessions may be dropped during a switchover. They should reestablish themselves automatically just as they would following a session drop for any other reason.

Switchovers on OC-3/OC-3 NI-2 cards take slightly longer to complete than switchovers on other types of NI-2 card. If PPP sessions are set to time out in 60 seconds session drops will occur more frequently on OC-3/OC-3 NI-2 cards than on other types of NI-2.

Workaround: None. [CSCds13589]

Peer NI2 Will Reboot When Fuse Blows on Other NI-2

Under certain NI-2 card power fail conditions, the failing NI-2 card may reset the standby NI-2 card.

Workaround: None. [CSCdt90949]

Redundant NI-2s Reload When All DSL Ports Drop Simultaneously

This problem occurs in redundant systems fully populated with octal-port line cards and in which every port is attached to a single profile. Under these circumstances, when the profile is changed, the large volume of port status messages can keep the active card from communicating with the standby. When this happens the standby card resets the active card and takes activity. If the chassis is a Cisco 6160, the card taking activity will not be able to communicate with the line cards.

Workaround: Use a different profile for each quadrant of the chassis even if the four profiles contain the same information. This configuration prevents the flood of port status messages during profile changes. [CSCdu14280]

Autosync of Start-up Config Does Not Always Work

In redundant configurations, automatic synchronization of the startup configuration does not function reliably. The problem occurs most frequently with larger configuration files (as in the head node of a subtend tree or a system fully configured with octal,-port line cards), but because the problem is sporadic, the recommended workaround is to disable automatic configuration synchronization and to always manually synchronize the start-up configuration following configuration changes.

Workaround: Disable automatic configuration synchronization. Manually synchronize the start-up configuration following configuration changes. The command to do so is:

DSLAM#secondary sync config

[CSCdu15630]

8xDMT Caveats

677 3.0 Cat1 Down Trains Lower Without Noise Than With

Laboratory testing of the 8xDMT with the Cisco 677, with the 3.0 ADI firmware, may yield inferior results when noise is injected only at the CPE end of the link on Category 1 loops. Testing with noise injected at both the CO and CPE ends of the link yields expected performance over Category 1 loops.

Workaround: Since noise at both the CO and CPE ends of the DSL link is a more accurate representation of an actual deployment, it is recommended that the customer only test with noise injected at both ends of the DSL link. [CSCdt32390]

G.Lite Loss of Frames, CPE Will not Train

When connecting a Cisco 678 CPE to an 8xDMT provisioned for G.lite:

At 15,000 feet, 26 AWG wire:

: There are loss of frames at provisioned margins of 0/0, 1/1, and 2/2 (up margin/down margin).

At 17,000 feet, 26 AWG wire:

: The CPE does not remain trained at provisioned margins of 0/0 and 1/1. There are loss of frames at a provisioned margin of 2/2.

Workaround: Do not use provisioned margins of 0/0, 1/1, or 2/2 when connecting a Cisco 678 CPE to an 8xDMT provisioned for G.lite when line lengths are 15,000 feet or greater. [CSCdt78613]

The 8xDMT Line Card Ignores Interleave Setting When Trained to 677

The interleave delay value specified in the DMT profile may not match the value reported by the port when training is complete.

Workaround: None. [CSCdu09877]

4xSDSL Caveats

Profile Lists Rates of 1552, 2064, 2320 & Allows Their Use

The maximum data rate for the 4xSDSL is 1168 kbps. Rates of 1552, 2064, and 2320 are configurable but should not be used.

Workaround: None. [CSCdt76848]

Open Caveats—Release 12.1(6)DA2

Active LED on 8xDMT Cards Stays on When NI-2 Communication Lost

When the 8xDMT or line card establishes initial communication with the NI-2 card, the active LED on the line card face plate will be illuminated. If communication with the NI-2 card is lost subsequent to this, through a reload command or removal of the NI-2 card, the active LED will remain illuminated.

Workaround: None. [CSCdu12337]

Open Caveats—Release 12.1(6)DA

The caveats in this section apply only to Cisco IOS Release 12.1(6)DA. Other open caveats that apply to this Release 12.1(6)DA are described in the "Open CaveatsRelease 12.1(5)DA" section, "Open CaveatsRelease 12.1(4)DA" section, the "Open CaveatsRelease 12.1(3)DA" section, the "Open CaveatsRelease 12.1(2)DA" section, "Open CaveatsRelease 12.1(1)DA1" section, and the "Open CaveatsRelease 12.1(1)DA" section.

4xFlexi-DMT Caveats

677 2.x Reports Incorrect Up Margin for Cat1 Loop midCSA6

An erroneous upstream margin value is reported by 4xflexi-DMT when running Cat1 rate-adaptive tests with loop midCSA6. The value observed is 63.5. This occurs only when connected to a Cisco 677 CPE.

Workaround: None. [CSCdt32428]

c677 w/3.0 FW and Cat1 Loop Anomaly with Injected Noise

Layer 1 Cat1 adaptive rate test on ansi#7 loop showed higher downstream rates when noise was injected on CPE side instead of CO side. Injecting noise at both sides simultaneously also yielded better downstream results than for injecting on CO only.

Workaround: None. However, injecting noise at both ends at the same time will give more reasonable results. [CSCdt32401]

Alcatel-Based CPEs Training Odd at High Bit Rates, Interleaved

When using 4xflexi-DMT and a configured downstream max bit rate (interleaved) above 7616 kbps, Alcatel chip set based CPE (for example, Westell, Efficient, Speedtouch, etc) train to a lower bit rate and not always as expected.

Workaround: You can get the desired max rate if you use max rates less than 7616. [CSCdt16528]

Degraded Downstream Bit Rates at Different Lengths from 12.1(5)DA

When flexi-DMT line card is used with Efficient 5262 CPEs, the trained bitrates performance is not comparable to that of Cisco IOS Release 12.1(5)DA.

Workaround: Experiment with different settings of margin, check-bytes, and interleave delay to achieve the bit rate desired.[CSCds81181]

4xFlexi-DMT Unable to Train to Maximum. Rates with 677 CPEs

Configured maximum bit rates on a 4xflexi-DMT are not attainable when trained against a Cisco 677 CPE.

Workaround: None available. [CSCds76055]

Next: Low Reported Margin for 677 Firmware 2.2

A 4xflexi-DMT reports very low downstream margins when connected to a Cisco 677 CPE.

Workaround: None. [CSCdt28665]

GSI reports that the margin number being reported by the ADI-based CPE and communicated over the EOC channel does not include coding gain and thus appears as a low value. We are not anticipating any new developments on the Cisco 677 CPE front, therefore we are simply documenting this behavior.

Note Because of this behavior, Cisco cannot recommend SNR-based line quality monitoring when using a Cisco 677 CPE.

Running-config Incorrectly Marked as Changed

When one or more 4xflexis are provisioned in a system, a reload command will always result in the following query. System configuration has been modified. Save? [yes/no]:

Workaround: None. [CSCdt41089]

ST Pro in G.lite Mode Fails Several Annex D Test Loops

The 4xflexi fails layer 1 G.lite mode-adaptive rate tests.

Workaround: None. [CSCdt34800]

Takes More Time Than Downshift Interval to Meet Rate Adaptation

When rate-adaptation mode is enabled, the configured downshift interval for the downstream direction does not appear accurate.

Workaround: None. [CSCdt03800]

Westell HEC Errors, Frame Loss for Low Margins (0 to 3)

Downstream HEC errors and uncorrected super frames may occur when the 4xflexi-DMT line card is used with some Alcatel chipset-based CPE if a low target/configured margin is used.

Workaround: Set the configured margin to values higher than 3. The default of 6 is fine. [CSCdt32319]

MIB Caveats

Creation of snmpNotifyFilterProfileTable Incorrect—Can Cause Reboot

If rowstatus is set to createAndWait, no indexes are created to allow for configuration of variables in the table. If the table is walked, it will loop. The DSLAM can reload when trying to set any of the variables.

Workaround: None. [CSCds70225]

Miscellaneous Caveats

Filesys space Error When 1st Write Mem After Reload

Filesys space error is displayed when performing the first write mem following a reload.

Workaround: None. The next write mem issued will not produce an error. [CSCdt27044]

8xDMT Caveats

677 2.x Reports Incorrect Up Margin for Test Loop midCSA6

Using a Cisco 677 CPE that contains ADI chipset version 2.x, with an 8xDMT, the upstream margin is sometimes reported erroneously as -63.5 db.

Workaround: None. [CSCdt32415]

8xDMT to 677 3.0 CSF, USF, and Margin Perf May Differ from Flexi

When connecting the 8xDMT to a Cisco 677 CPE running version 3.0 firmware, the corrected and uncorrected superframe counts and margin performance may differ from that of the flexi card.

Workaround: None. [CSCdt23551]

Actual Margin < Provisioned for G.lite and DMT

This behavior occurs with a Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMT line card. A port that has been configured for a specific minimum margin may, on occasion, train with a margin 1 or 2 dB less than the configured minimum margin.

Workaround: The configured bit rate may be lowered slightly to allow a greater margin when trained, or the minimum margin downshift feature may be enabled, causing the port to retrain if the margin is below a specified minimum over a specified interval of time. [CSCds51398]

Alcatel Modems: no train when ds<=1024 and Trellis=disabled

When you use a Cisco 827, Efficient 5260, or Westel CPEs with the 8xDMT and provisioned with trellis encoding disabled, the downstream bit rates will not train below 1024 kbps.

Workaround: Enable trellis encoding when using one of these CPEs with the 8xDMT. [CSCds56646]

Downstream Margin not Reported with Cisco 677 CPE (ADI ver. 2.2)

When you connect the 8xDMT to a Cisco 677 CPE running ADI firmware version 2.2, the downstream margin reported is not valid. This ADI firmware version does not support the EOC mechanism used for reporting downstream margin.

Workaround: None. [CSCdt24787]

Downstream USF in Fast Path with Cisco 678

When you use a Cisco 678 CPE with the 8xDMTs, if the port is configured for fast latency there may be uncorrected downstream superframes on longer loops.

Workaround: When configuring the port, use interleaved latency if possible. [CSCdt17039]

Fast Latency Will not Train Downstream < 64 kbps

When you configure a Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMTs, configuring fast latency operation with maximum bit rates set to 32 kbps in both upstream and downstream directions causes the port to fail to train.

Workaround: Configure the maximum downstream bit rate to a value greater than 32 kbps. [CSCds69455]

Increased Corrected and Uncorrected SF Counts US with 5262 CPE

This occurs when you connect a Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMTs to an Efficient 5262 CPE. If a port is configured for Interleaved Latency mode when using the 5262, and loop length is in the 8 to 9 Kft range, the upstream data path may occasionally accumulate higher corrected and uncorrected superframe counts.

Workaround: None. [CSCdt23206]

Increased Corrected and Uncorrected SF Counts with Westell CPE

A Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMT s, configured for fast latency mode with a Westell (AME 1.4.1) CPE, will occasionally accumulate higher corrected and uncorrected superframe counts in both upstream and downstream data.

Workaround: None. [CSCdt24780]

LOF, LOS, LPR may be Reported Inaccurately

Performance data, including both totals and interval information, is collected for display through the SNMP. The following statistics may be reported inaccurately: near-end LOF (loss of frame), near-end LOS (loss of sync), far-end LOF, far-end LOS, and far-end LPR (loss of power).

Workaround: None. [CSCdt28479]

The Margin Reported with 678 CPE may Differ from Actual

When a Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMT s is connected to a Cisco 678 CPE, the margin value reported may be slightly greater than the actual trained margin.

Workaround: None. [CSCdt04507]

Unable to Obtain Requested Int. Delay with /677

When you connect to a Cisco 677 CPE, the actual downstream interleave delay may be slightly greater than the configured delay, when less than 16 ms.

Workaround: Provision the downstream interleave delay to a minimum of 16 ms. [CSCdt16909]

Upstream Margin not Accurate for Alcatel-based CPE and Cisco 678

The DSLAM reports the upstream margin of -63 dB or 63 dB (depending on the DSP firmware version) when operating with Alcatel-based CPEs (tested with Efficient 5260, 5262, and Cisco 827) or with the GSI-based Cisco 678 CPE that is provisioned for fast latency. Due to this problem, the rate-adaptation feature implemented in 12.1.6(DA) will not work properly when the upstream margin is not reported correctly.

This happens when trellis encoding is disabled and the operating mode is in T1.413.

Workaround: For Alcatel-based CPEs, enable trellis encoding when configuring the port. Fast latency is not supported when operating with a GSI-based Cisco 678 CPE. [CSCdt18971]

Variable Train Rates with 677 CPE (12.5 to 13.0 kft)

This behavior occurs with a Cisco 6015, Cisco 6160, or Cisco 6260 DSLAM equipped with the 8xDMTs, connected to a Cisco 677 CPE running firmware 3.0. The train rates are not consistent over reach, with multiple trains on the same port occasionally exhibiting variations in rate.

Workaround: None. [CSCdt24795]

Proxy Element Supporting the Cisco DSL CPE Automated Configuration Solution

CIB Save FAILED Message Displayed During NI Boot

During bootup of the DSLAM, a message is displayed that the CIB table cannot be saved to Flash. This is due to SNMP saving configuration information such as the reboot count.

Workaround: None. During bootup, the CIB table should be prevented from getting saved to the Flash. This is the desired behavior. Additional logic will be added to Cisco IOS 12.1(7)DA to prevent the unexpected save request.[CSCdt27246]

DSLAM Lockup with Multiple Slot/no Slot Commands

With CDCM attached, the DSLAM will lockup when multiple slot/no slot commands are issued.

Workaround: Deactivate CDCM prior to issuing slot/no slot commands. [CSCds90782]

Open Caveats—Release 12.1(5)DA

The caveats in this section apply only to Cisco IOS Release 12.1(5)DA. Other open caveats that apply to this Release 12.1(5)DA are described in the "Open CaveatsRelease 12.1(4)DA" section, the "Open CaveatsRelease 12.1(3)DA" section, the "Open CaveatsRelease 12.1(2)DA" section, "Open CaveatsRelease 12.1(1)DA1" section, and the "Open CaveatsRelease 12.1(1)DA" section.

4xFlexi Caveats

FastPath Working on G.lite

With the 4xflexi-DMT configured, the line is able to train up in fast path when in G.lite mode. This was observed with CPEs capable of G.Lite and fast path.

Workaround: Do not enable fast path when in G.Lite operating mode. [CSCds52496]

BERT Test Interrupts Upstream Traffic on the Other Card Ports

With the 4xflexi configured, other ports on the same line card are unable to pass upstream traffic when one of the ports on the line card is running a BERT test. Once the BERT test completes, the upstream traffic on other ports resumes.

Workaround: Do not initiate the BERT test if there is traffic running on other ports on the same line card. [CSCds73901]

Downstream Frame Loss When Check-bytes Less Than 6 with 5262/Interleave

With the 4xflexi-DMT configured for interleave path operation and an FEC value of less than 6, a number of Uncorrected Superframes and thus lost data frames (in the order of 1 in 10,000) in the downstream direction can occur on certain loops. This was observed with line card dsp image version K.29 while connected to an Efficient 5262 CPE.

Workaround: Set an FEC value of 6 or greater.[CSCds76533]

MIB Caveats

adslAturCurrOutputPwr Returns Incorrect Value on CAP Cards

Setting adslAtuc/AturCurrOutputPwr in the ADSL-LINE-MIB returns an incorrect value for short loops (< 2000 feet). When Transmit Power reports a negative value, the SNMP query to these variables returns an incorrect value in the response.

Workaround: Use CLI. [CSCds48052]

adslAtucCurrStatus Returns Incorrect Value for CAP and DMT

Setting adslAtucCurrStatus returns an incorrect status for CAP and DMT.

Workaround: Use CLI. [CSCds80863]

adslAturCurrStatus Returns Incorrect Value for DMT

Setting adslAturCurrStatus returns an incorrect status for DMT line cards.

Workaround: Use CLI. [CSCds81122]

cAdslAtuc/AturCapCurrRxGain Returns Invalid Values

Setting cAdslAtuc/AturCapCurrRxGain in the CISCO-ADSL-CAP-LINE-MIB returns an incorrect value for short loops ( < 2000 feet). When Receiver Gain reports a negative value, the SNMP query to these variables returns an incorrect value in the response.

Workaround: Use CLI. [CSCds48063]

Sonetpath(50) Missing from ifTable

The ifType with a value sonetPath(50) is missing from the ifTable. The ifIndex of MIB variables in sonetPathCurrent and sonetPathInterval tables are mapped to the sonet(39) ifType. The ifIndex of these variables should map to a sonetPath(50) ifType, which is not present in the ifTable.

Workaround: None. [CSCds59504]

Miscellaneous Caveats

SNMP set cmd with string param (Cisco IOS CLI) Causes System Failure

Executing an SNMP Set/Inform request with string or ip address parameter through Cisco IOS CLI causes an unexpected DSLAM reload. For example:

snmp set v3 ip xxx.xxx.xxx.xxx community pef context pef oid cdcpCibOwner.0 string pef

Other parameters, for example, integer work fine.

Workaround: Use other tools to set objects that require a string or ip address input. [CSCds81489]

MPLS Caveats

NI-2 Reload under Destructive VPN Routes Update Test

Running a destructive test by sending more than 30,000 VPN routes across the MPLS network causes the NI-2 to reload.

Workaround: Configure the maximum routes for each VPN to prevent the CE devices or remote PE routers from sending the VPN routes more than it should. [CSCds51409]

Open Caveats—Release 12.1(4)DA

The caveats in this section apply only to Cisco IOS Release 12.1(4)DA. Other open caveats that apply to Release 12.1(4)DA are described in the "Open CaveatsRelease 12.1(3)DA" section, "Open CaveatsRelease 12.1(2)DA" section, "Open CaveatsRelease 12.1(1)DA1" section, and "Open CaveatsRelease 12.1(1)DA" section.

IMA Caveats

E1 IMA Group Bandwidth is Limited to Six Links

E1 IMA group bandwidth is limited to six links. Cell loss will occur if an IMA group has more than six links and more than 12 Mbps are sent through it.

The loss only occurs on groups with greater than six E1 links. All eight links can be used in different groups, however (for example, six in group 0, two in group 1). This does not affect T1 operation (that is, all eight T1 links may be used in a group).

Workaround: Do not configure an IMA group with more than six E1 links. [CSCdr45623]

Cell Loss May Occur on Subscriber Ports If Trunk is Down

Cell loss on the subscriber (DSL) ports may be seen if you select an IMA group and the trunk is down. The loss only happens if the trunk IMA group is down. [CSCds02810]

Workarounds:

Shutdown the trunk IMA group.
Bring the trunk IMA group up (by troubleshooting).
Reassign the trunk to another interface (changes network topology).

MPLS VPN Mapping of RFC 1483 Routed Sessions

Attempting to Create a Circuit on VCI 3 or 4 is Allowed by the CLI

The CLI incorrectly accepts configuration input on reserved VCIs 3 and 4 through the atm pvc command.

Example:

DSLAM#configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
DSLAM(config)#interface atm1/3
DSLAM(config-if)#atm pvc 1 4 interface atm0/0 0 120
DSLAM(config-if)#
DSLAM#show running interface atm1/3
Building configuration...
 
Current configuration : 62 bytes
!
interface ATM1/3
 no ip address
 no atm ilmi-keepalive
end                            
 
VC-Table
ATM1/3        1     4      PVC     ATM0/0        0     120   SNAP   DOWN

If you attempt to create virtual channel connections (VCC) 1/3 or 1/4 using the pvc command an error message appears.


DSLAM(config)#interface atm1/3
DSLAM(config-if)#pvc 1/3
%ATM: Invalid VCI of 3 requested: (ATM1/3): Not creating vc:1:3
DSLAM(config-if)#pvc 1/4
%ATM: Invalid VCI of 4 requested: (ATM1/3): Not creating vc:1:4
DSLAM(config-if)#pvc 1/5
DSLAM(config-if-atm-vc)#

Workaround: Do not attempt to configure VCIs 3 and 4. [CSCdr99296]

Changing Line Coding or Line Card Type Causes Duplicate VPN References in show ip vrf Output

If you swap a line card or change the line coding on a DSL interface, the show ip vrf command may display duplicate references to a VRF instance.

Workaround: Do not swap line cards with a different type. Do not change line coding on an active interface. [CSCds29146]

Configuring a Large Number of Subinterfaces Causes Service Loss

Configuring more than 32 PVP subinterfaces on the trunk interface of the NI-2 may cause a loss of connectivity to remote peer MPLS routers.

Workaround: Currently, there is no workaround. [CSCdr88865]

Configuring IP MTU or Physical MTU Size has no Effect on Fragmentation

No fragmentation occurs at the IP layer.

Workaround: There is no workaround. The default MTU settings work fine for Layer 3 services on the DSLAM. [CSCds15469]

debug atm error Command Generates Invalid Error Messages

Using the debug atm error command generates the following invalid error message:

"ATM(ATM0/1):point-to-point interface does not have a VCD"

The message appears when there is no PVC defined on the atm0/1 physical interface.

Workaround: This message can be ignored. Configuring a point-to-point PVC on the physical uplink interface is unnecessary for MPLS functionality. [CSCdr97541]

DSL Interface Resets Under Heavy Stress

DSL interfaces reset when CPU usage approaches 100 percent for an extended period of time.

Workaround: Using the show proc cpu command, identify the CPU hogging interfaces. Move those interfaces to a less taxed IP DSL Switch. verify that the CPU usage exceeds 95 percent. Then, using the show interface accounting command, take successive samplings of packet counters to identify the interfaces with the highest traffic rates. Once these interfaces are identified, move them to a DSLAM with lower CPU usage rates. [CSCds23638]

Erroneous debug atm Error Message When Creating a PVC or MPLS-Switched Interface

Output from the atm debug command returns misleading error messages while configuring a point-to-point PVC or MPLS-switched interface.

The following erroneous message is logged:

c6100_atm_add_vc_to_swsb: Error adding VC to vc-list

Workaround: This message can be ignored. The VC creation has taken place. [CSCds15732]

Interface Site-of-Origin Tagging is Not Working

Site-of-Origin tagging does not work when associated with the ip vrf sitemap on customer edge (CE) interfaces.

Workaround: Currently, there is no workaround. [CSCdr93090]

IP DSL Switch Accepts Invalid ATM Syntax on DSL Interfaces

The ATM command syntax atm pvc vpi vci is not parsed correctly by the DSLAM. [CSCds05174]

Workaround: Use the following syntax:

atm pvc vpi vci interface atm [slot/port] vpi vci

Maximum Number of Interior Gateway Protocol (IGP) Routes Supported on the IP DSL Switch is Limited to 3500

The maximum number of Interior Gateway Protocol (IGP) routes on the DSLAM is limited to 3500.

Workaround: Currently, there is no workaround. [CSCdr89959]

Outbound Access Lists Do Not Work

Outbound access- lists on uplink and CE interfaces are ignored.

Workaround: Currently, there is no workaround. [CSCds01820]

Packet Sizes of Less Than 512 Bytes Result in Reduced Aggregate Throughput in the NI-2

Aggregate throughput in NI-2s running Layer 3 services is poor when average packet sizes are less than 512 bytes.

Workaround: Currently, there is no workaround. [CSCdr92284]

Sending More than 12,500 RIP Routes in a Single VPN Causes High CPU Usage

With a large number of VPN routes in a single VPN domain, the CPU may approach 100 percent for too long when more than 500 routes expire at the same time. This causes system instability.

Workaround: Segment the network into multiple VPN domains or use the maximum route # command to limit the number of VPN routes into the network. For example: [CSCds27853]

 ip vrf vpn1
  rd 100:1
  route-target export 100:1
  route-target import 100:1
  maximum routes max number of routes Threshold value%

Static IP Routing Tables Break After Line Coding Change

Static routes are not applied to a DSL interface if you change its line coding.

Workaround: Remove all static IP routes using the no ip route and no ip route vrf commands before you change line coding. Recreate the static routes after the line coding is changed. [CSCds38810]

System Generates a MALLOCFAIL Error if You Disable Cisco Express Forwarding While Traffic is Running

Disabling Cisco Express Forwarding (CEF) while MPLS traffic is being passed can destabilize the IP DSL Switch.

Workaround: CEF is a prerequisite to running MPLS. It should never be disabled while MPLS traffic is being passed. Leave CEF enabled. [CSCds29151]

System Generates an Incorrect Error Message When Attempting to Remove VP-tunnel

If you attempt to remove a VP-tunnel interface on the atm0/1 you receive the following incorrect error message:

Example:

DSLAM(config)#interface atm0/1.60
DSLAM(config-subif)#no tag-switching atm vp-tunnel 60
%Invalid tag vpi range for selected VP-tunnel
Hardware does not support the specified VPI, 1

Workaround: None. This message can be ignored. [CSCdr89031]

System Generates Unknown Protocol, VRF Error Message Reference During NI-2 Reload

When you create a VPN routing and forwarding instance (VRF) on systems with an older bootflash image, the following message may appear:

% Can't create VRF <VRF-ID>

This is message is generated because of an outdated bootflash image.

Workaround: The message can be ignored. It has no impact on MPLS VPN functionality. [CSCdr89698]

Open Caveats—Release 12.1(3)DA

The caveats in this section apply only to Cisco IOS Release 12.1(3)DA. Other open caveats that apply to Release 12.1(3)DA are described in the "Open CaveatsRelease 12.1(2)DA" section, "Open CaveatsRelease 12.1(1)DA1" section, and "Open CaveatsRelease 12.1(1)DA" section.

4xDMT over ISDN Caveats

This section describes problems with the 4xDMT over ISDN.

With Trellis Enabled, Errored Seconds are High

Enabling DMT trellis line coding on the 4xDMT over ISDN (ADI-based) can decrease link reliability by causing problems passing data due to errored second events and data corruption in the downstream and upstream directions. ADI/Aware has a known problem where high errored seconds may result if trellis encoding is turned on. This problem is not scheduled to be fixed in the 2.0 release, which is currently where the 4xDMT over ISDN code is branched.

On the 4xDMT over ISDN, having trellis encoding enabled will decrease link reliability by causing problems training or passing data. There is no way to accurately predict when it will be reliable. Therefore, trellis encoding is disabled by default on the NI-2 to ensure that it is not accidentally enabled with the 4xDMT over ISDN.

Workaround: DMT trellis line coding is not enabled by default on the NI-2, because it is not supported on the 4xDMT over ISDN. [CSCdr92590]

Some Trains in Quick Mode will have High Errored Seconds

This problem occurs at long loop lengths (11500 feet and above) when using the quick training mode of the 4xDMT over ISDN. It can decrease link reliability by causing problems passing data due to errored second events and data corruption in the downstream direction.

On the 4xDMT over ISDN, using the quick training mode at long loop lengths (3600 m and above) will decrease link reliability. The link will train at a higher data rate, but at lower downstream margins (:for example, 1 or 2 dB) rather than training at a lower data rate and maintaining the requested downstream margin (:for example, 6 dB). This can decrease link reliability by causing errored second events and data corruption in the downstream direction due to the acquired lower downstream margins.

Workaround: Use the standard training mode to train up the link at long loop lengths (11500 feet and above). [CSCdr93540]

Reported Margins do not Match Selected Margins

This problem is identical to the problem described in DDTS report CSCdm89068, which describes the same symptoms that appear when using the 4xDMT.

In some situations, the reported (actual) margins may be higher than the configured values. Depending on the port configuration, there are effective minimum end-to-end margin values for combinations of parameters including:

Training mode
Framing mode
Interleaving delay
Trellis encoding
Number of FEC checkbytes

Workaround: If you configure the margins for a port below these effective minimum margins, the system accepts the margin configuration, but the margin values that are actually used and reported are the effective minimum end-to-end values. The actual margins are essentially clipped at these effective minimum values. [CSCdr93546]

The effective minimum margins for the downstream and upstream directions are given in the following tables:

Train Mode	Frame Mode	Intlv >0	Trellis Encoding	Ckbytes >0	Downstream Bit Rate Range (kbps)
Train Mode	Frame Mode	Intlv >0	Trellis Encoding	Ckbytes >0	32 to 640	672 to 768	800 to 3040	3072 to 6112	6144 to 8032
X	X	X	X	No	0	0	0	0	0
Quick	0,1	No	Disabled	Yes	0	5	5	4	5
Quick	0,1	No	Enabled	Yes	0	6	6	5	6
Quick	0,1	Yes	Disabled	Yes	0	6	6	5	6
Quick	0,1	Yes	Enabled	Yes	0	7	7	6	7
Quick	2,3	No	Disabled	Yes	0	0	5	4	5
Quick	2,3	No	Enabled	Yes	0	0	6	5	6
Quick	2,3	Yes	Disabled	Yes	0	0	6	5	6
Quick	2,3	Yes	Enabled	Yes	0	0	7	6	7
Stnd.	0,1	No	Disabled	Yes	2	2	2	1	2
Stnd.	0,1	No	Enabled	Yes	3	3	3	2	3
Stnd.	0,1	Yes	Disabled	Yes	3	3	3	2	3
Stnd.	0,1	Yes	Enabled	Yes	4	4	4	3	4
Stnd.	2,3	No	Disabled	Yes	2	2	2	1	2
Stnd.	2,3	No	Enabled	Yes	3	3	3	2	3
Stnd.	2,3	Yes	Disabled	Yes	3	3	3	2	3
Stnd.	2,3	Yes	Enabled	Yes	4	4	4	3	4

Train Mode	Frame Mode	Intlv >0	Trellis Encoding	Ckbytes >0	Upstream Bit Rate Range (kbps)
Train Mode	Frame Mode	Intlv >0	Trellis Encoding	Ckbytes >0	32 to 640	672 to 768	800 to 864
X	X	X	X	No	0	0	0
Quick	0,1	No	Disabled	Yes	0	5	5
Quick	0,1	No	Enabled	Yes	0	6	6
Quick	0,1	Yes	Disabled	Yes	0	6	6
Quick	0,1	Yes	Enabled	Yes	0	7	7
Quick	2,3	No	Disabled	Yes	0	0	5
Quick	2,3	No	Enabled	Yes	0	0	6
Quick	2,3	Yes	Disabled	Yes	0	0	6
Quick	2,3	Yes	Enabled	Yes	0	0	7
Stnd.	0,1	No	Disabled	Yes	2	2	2
Stnd.	0,1	No	Enabled	Yes	3	3	3
Stnd.	0,1	Yes	Disabled	Yes	3	3	3
Stnd.	0,1	Yes	Enabled	Yes	4	4	4
Stnd.	2,3	No	Disabled	Yes	2	2	2
Stnd.	2,3	No	Enabled	Yes	3	3	3
Stnd.	2,3	Yes	Disabled	Yes	3	3	3
Stnd.	2,3	Yes	Enabled	Yes	4	4	4

8xIDSL Line Card Caveats

This section describes problems with the8xIDSL.

Clocking Problem

When you enter the show controllers idsl card/port command after a reload/reboot, an excessive number of network clock losses is reported. This is normal because the network interface module is being reinitialized and network clocking has yet to be established.[CSCdr91534]

Workaround: None.

Miscellaneous

The following are miscellaneous problems with this release.

The IDSL-associated atm-p interface input counters are not being updated. The per-VC counters are updating properly, but interface-level counters are not.

Workaround: Use the following CLI commands that display counters:

show interface idsl card/port
show atm vc traffic interface atm-p card/port [CSCdr90732]

Loopback and Test Caveats

The following are loopback and test problems with this release.

When you configure an interface for loopback diagnostic operation and this configuration becomes part of the startup-config, the interface might initialize in the down state after DSLAM reload/reboot.

: Workaround: Enter the clear interface idsl card/port command to reinitialize the interface. [CSCds01533]

On rare occasions, if you issue the remote loopback command, the interface goes down and remains down. You can verify this failure mode by issuing the show controllers IDSL slot/port command and confirming that the C/I state is not 0xF3.

: Workaround: Remove and reissue the loopback remote command. [CSCdr62772]

On rare occasions, an 8xIDSL BERT test will report bit errors and multiple sync losses.

: Workaround: Remove and reissue the loopback remote command. [CSCdr63116]

ATM Traffic Shaping and Policing not Functioning

On the Cisco 6160 DSLAM equipped with 8xIDSLs, ATM traffic shaping and policing might not function properly. Shaping is enabled on Frame- Relay and PPP PVCs and soft PVCs with the UPC drop option of the frame-relay pvc, frame-relay soft-vc, ppp pvc, or ppp soft-vc commands. Policing without traffic shaping (that is, tagging of cells exceeding the ATM traffic descriptor parameters) is enabled with the upc tag option of the above four commands.

Specifically, the following problems are known to exist:

Shaping and policing do not work if the ATM SCR is less than 3 kbps. The behavior is that all traffic is passed and no cells are dropped or tagged.
Shaping and policing at higher rates does function, but the rate to which shaping and policing occurs might not match the configured rate.

There are no workarounds to the above problems.[CSCdr70582]

Two NI-2 Cards have the same ATM Address

When an NI-2 card is shipped, the ATM address is blank. When that NI-2 card is provisioned in a chassis, an ATM address is generated using the MAC address of the chassis. If that NI-2 is moved to another chassis and a new NI-2 card is installed in the original chassis, the two NI-2 cards will have the same ATM address. [CSCds26889]

Workaround:

1. Issue the command show hardware chassis and note the MAC address from the IO card EEPROM field.

2. Issue the command show atm address and note the ATM address.

3. Issue the command configure terminal.

4. Issue the command atm addr <enter the first 7 bytes from the original address><enter MAC addr><enter MAC addr again>.00.'

5. Enter no atm address (original atm address).

6. And finally issue reload.

Open Caveats—Release 12.1(2)DA

The caveats in this section apply only to Cisco IOS Release 12.1(2)DA. Other open caveats that apply to Release 12.1(2)DA are described in the "Open CaveatsRelease 12.1(1)DA1" section, and in the "Open CaveatsRelease 12.1(1)DA" section.

IDSL Line Card Caveats

Frame Relay Configuration Caveats

The NI-2 cannot terminate IDSL FRF.8 service transparent PVCs. The system accepts commands of the form frame-relay pvc... service transparent interface ATM0/0... encap aal5mux ip, but the PVC is not properly terminated. Only service translation PVCs can be terminated by the NI-2 with an ATM encapsulation of aal5snap. Note that service transparent PVCs can be terminated via any physical ATM interface, however. [CSCdr16342]
If you issue the shutdown command on a Cisco 6160 IDSL interface that is attached to a Motorola Vanguard 65, and Frame Relay encapsulation is configured on both devices, a subsequent no shutdown command might not bring the Motorola Vanguard 65 interface back up. This problem does not occur if PPP encapsulation is configured.

: Workaround: Power cycle the Motorola Vanguard 65. [CSCdr32145]

IDL 144 kbps mode is currently disabled for IDSL interfaces. It will be enabled in a future release. At that time, the command dsl profile idsl bitrate 144 idl will be enabled. Note that 144 kbps non-IDL mode is supported. [CSCdp63832]
If you configure Frame Relay encapsulation on an IDSL interface, the interface does not pass Appletalk traffic.

: Workaround: For an End-to-End AppleTalk routing, the service type for a FR IWF pvc should be configured as 'transparent' on the DSLAM. The terminating end encapsulation on the ATM side should be 'aal5nlpid' for the pvc. [CSCdp70130]

Interface State Caveat

On rare occasions, the results of the command show interface idsl slot/port indicate that an interface is up when it is in fact down. You can verify this failure mode by issuing the command show controllers idsl slot/port and confirming that the "C/I" state is not "0xF3".

Workaround: Issue the shutdown command followed by the no shutdown command on the IDSL interface. [CSCdr23602]

Status Display Caveats

The show dsl interface idsl slot/port command displays excessive HDLC layer performance statistics errors if the line interface cable to the CPE is pulled. This is an aspect of the implementation and does not indicate negative performance. [CSCdr43144]

Loopback and Test Caveats

If you issue the no loopback command on an 8xIDSL interface on which a loopback command has not previously been issued, the interface goes down and back up. This causes a momentary interruption of service for the affected interface. [CSCdr35443]
A Cabletron Flowpoint 144F-16 IDSL router cannot be placed into remote loopback when it is operating in 144 kbps mode when you issue the loopback remote command on the Cisco 6160. You can put the Cabletron Flowpoint 144-16 IDSL router into remote loopback if it is operating at 64 kbps or 128 kbps. There is no workaround. [CSCdr58260]
On rare occasions, if you issue the loopback remote command and then issue the no loopback remote command, the interface goes down and remains down. You can verify this failure mode by issuing the show controllers idsl slot/port command and confirming that the "C/I" state is not "0xF3".

: Workaround: Disable and re-enable the interface using the shutdown command followed by the no shutdown command. [CSCdr60773]

Bit error rate testing (BERT) runs continuously if the Cisco 6160 is reloaded while BERT is in progress.

: Workaround: To abort BERT before reloading the Cisco 6160, use the dsl test idsl slot/port abort command. [CSCdr62092]

8xIDSL interfaces that are configured with Frame Relay encapsulation and LMI experience periodic interface state transitions if you issue the loopback remote command. This results in the periodic printing of LINK, LINEPROTO and possibly alarm transition messages to the console until you disable the loopback using the no loopback remote command. The interface transitions and the resulting console messages do not affect BERT tests and are completely benign. [CSCdr62656]
On rare occasions, if you issue the loopback remote command, the interface goes down and remains down. You can verify this failure mode by issuing the show controllers idsl slot/port command and confirming that the "C/I" state is not "0xF3".

: Workaround: Use the no loopback remote command to turn the remote loopback off, then use the loopback remote command to turn it on again. [CSCdr62772]

On rare occasions, an8xIDSL bit error rate test (BERT) reports bit errors and multiple sync losses.

: Workaround: Use the no loopback remote command to turn the remote loopback off, then use the loopback remote command to turn it on again. [CSCdr63116]

SDSL Line Card Caveat

In a fully loaded chassis after hours of normal operation, one or more ports of a line card begins to have trickle CRC errors at medium to high data rates. The problem resolves itself upon retrain. [CSCdr46450]

4xFlexi Caveats

Trellis Encoding Disabled Can Cause Errors with 4xFlexi-DMTs

We recommend that trellis encoding be enabled in this release for the 4xflexi-DMT.

Note We recommend that you disable trellis encoding for 4xDMTs. See the "Incorrect Margin Value Reported for Some Cisco 678 Connections" section.

Interleave Delay of Zero Can Cause Errors with 4xFlexi-DMTs

We recommend that the interleave configuration be set to 1000 microseconds or greater for optimal reach performance. [CSCdr27700]

Line Card Alarms Not Cleared When a 4xFlexi Changes Type

When you change a 4xflexi card mode from DMT to CAP or CAP to DMT, the active alarms are not cleared.

Workaround: Before you change the line card provisioned mode, clear the existing alarms by issuing shutdown commands to all ports with active alarms. [CSCdr13611]

Incorrect Margin Value Reported for Some Cisco 678 Connections

When the DSLAM connects to a Cisco 678 CPE running a software release older than CBOS Release 2.3.5, the downstream margin reported is half of the actual margin.

Open Caveats—Release 12.1(1)DA1

The caveats in this section apply only to Cisco IOS Release 12.1(1)DA1. Other open caveats that apply to Release 12.1(1)DA1 are described in the "Open CaveatsRelease 12.1(1)DA" section.

Ethernet Port Cannot Recover Itself

When IP routing is enabled on an NI-2 DSLAM, the Ethernet port 0/0 does not come back up by itself after it experiences a problem. To reenable the port, enter the shut command in interface configuration mode, then enter the no shut command.

Workarounds: Do one of the following:

Use IP host-routing to establish connectivity over the Ethernet interface.
Turn off keep-alives on the Ethernet interface. If you do this, however, the system does not tell you when the Ethernet line goes down. [CSCdr56410]

NI-2 Cannot Recover if Bootflash Image Is Deleted

Xmodem, which resides in rommon on the NI-2 card, is not working. As a result, there is no way to recover if you delete the software image in bootflash and then try to reboot the system. In this case you must replace the NI-2.

Workaround: If you delete the image in bootflash, copy a new image into bootflash before shutting down or rebooting the system. [CSCdr22773]

Open Caveats—Release 12.1(1)DA

The caveats in this section apply only to Cisco IOS Release 12.1(1)DA.

Incorrect Margin Value Reported for Some Cisco 678 Connections

When the DSLAM connects to a Cisco 678 CPE running a software release older than CBOS Release 2.3.5, the downstream margin reported is half of the actual margin.

Maximum VCI Bits Value on Subscriber Ports Can Exceed 8

When autoconfiguration is enabled on a subscriber port, the DSLAM might set the maximum VCI bits for that port to an incorrect value that exceeds the maximum allowed value of 8. This depends upon a negotiated value with the connecting port on the adjacent node.

Workaround: Use the command no atm auto-configuration to disable autoconfiguration for subscriber ports. (Autoconfiguration is enabled by default.) [CSCdr18239]

DMT Ports Do Not Train Under Some FEC Check Bytes Settings

This applies to the 4xDMT. If a port is configured for standard training mode and the number of FEC check bytes for the downstream direction is configured for 2, 6, 10, or 14, then the port may have trouble staying trained if the loop conditions (such as loop length and noise levels) are favorable for training at a high bitrate. In other words, if the port would otherwise train at or above certain threshold bitrates in the downstream direction, and the FEC check bytes per codeword is set for 2, 6, 10, or 14, then the port continually retrains.

This problem does not occur if any of the following are true:

The port is configured for quick training mode
FEC check bytes per codeword for the downstream direction is configured for 0, 4, 8, 12, or 16.
The port trains below the threshold bit rate for the configured FEC check bytes per codeword (as listed in the table below).

The threshold bit rates at which this problem appears for each check byte per codeword setting are as follows:

FEC Check Bytes per Code Word	Threshold Downstream Bit Rate
0	Problem does not occur
2	8000 kbps
4	Problem does not occur
6	7872 kbps
8	Problem does not occur
10	7744 kbps
12	Problem does not occur
14	7616 kbps
16	Problem does not occur

Workarounds: Do one of the following:

Use the dmt training-mode command to configure the port for quick training mode.
Avoid setting the port to 2, 6, 10, or 14 check bytes per codeword in the downstream direction if it is possible that the line will train at or above 7616 Kbps. In this case, use settings of 0, 4, 8, 12, or 16 check bytes per codeword instead. Use the dmt check-bytes command to configure this parameter.
Use the dmt bitrate command to configure the maximum downstream bit rate below the threshold bit rates listed above, depending on the configured number of check bytes per codeword. [CSCdr16409]

Failed Ports Bring Down Healthy Ports on Same Line Card

When a hardware failure occurs on one or more of the ports on a 4xDMT, every other (otherwise healthy) port on the line card might be unable to train. Cisco strongly recommends that you replace or remove any line card with any defective ports. [CSCdp05249]

Problems Affecting Management Systems

These problems affect SNMP-based management systems, including CDM. You can avoid these problems by using the CLI.

You cannot set or change the subtend node ID using SNMP.

: Workaround: Use the CLI command subtend-id node# to set or change the subtend ID. [CSCdm93577]

Default ATM traffic descriptors cannot be deleted. When you attempt to delete them through the MIB, the MIB incorrectly indicates that the deletion is successful. [CSCdm89206]

4xDMT Caveats

Training Caveats

DMT lines sometimes train at 8096 kbps (downstream), although the maximum configurable bit rate is 8032 kbps. This occurs with Cisco 676 CPEs and might also occur with other vendors' Issue 1 equipment. [CSCdm66998]
DMT ports sometimes attempt to train (port LEDs blink) when no CPE is connected at the far end of the subscriber loop. This occurs when a neighboring port that is connected to CPE is training on a very short loop (less than 1000 feet). [CSCdm85326]
There may be no benefit in lowering configured end-to-end signal-to-noise ratio (SNR) margins below certain effective minimum values. Although the system allows margins to be configured below these effective minimums, the resulting bit rates and actual end-to-end margins might be the same as those realized when the margins are configured at the effective minimum values.

The effective minimum values depend upon

Training mode (standard or quick)
Bit rates to which the line trains
Whether interleaving is used

For the standard training mode, the effective minimum end-to-end margin is 1 to 3 dB. For the quick training mode, the effective minimum end-to-end margin is 4 to 6 dB. The effective margins might be better (that is, at the lower ends of the specified ranges) if interleaving is turned off or the bit rates fall between 6 Mbps and 3 Mbps. [CSCdm89068]

Trellis Encoding Can Cause Errors

We recommend that trellis encoding not be enabled in this release. (Trellis encoding is disabled by default.)

Although trellis encoding works reliably on most loops and most trains, it does not always work reliably. For some trains on some loops, enabling trellis encoding causes continuous errored seconds, as well as corrupted data in the upstream direction. It is not possible to predict the circumstances under which this happens. This problem will be fixed in a later release. [CSCdm73343]

Some Combinations of Framing Mode and Bit Rate Cause Data Corruption

Under certain combinations of upstream bit rate and overhead framing mode, no valid data flows upstream. When the line trains at these combinations, data going upstream is corrupted. Avoid these data rate and framing mode combinations:

Upstream Bit Rate	Overhead Framing Mode
32 kbps	Mode 1
64 kbps	Mode 2
96 kbps	Mode 3

Workaround: To avoid this data corruption problem, use these combinations of settings for the lower data rates.

Upstream Bit Rate	Overhead Framing Mode
32 kbps	Mode 3
64 kbps	Mode 1
96 kbps	Mode 1

[CSCdm66085]

Decreasing Interleaving Delay Increases Errored Seconds

When the DLSAM is connected to a Cisco 676 or any other CPE that is based on the ADI AD20MSP910 chipset (which was designed to support the older Issue 1 version of the T1.413 ADSL standard), you might experience higher than normal numbers of errored seconds, corrected superframes, and uncorrected superframes. This can occur when the interleaving delay is configured at values less than the maximum (16 ms). The errors can occur in both the upstream and downstream directions. This problem might be more apparent with short loop lengths (for example, those less than 4000 feet).

Workarounds: Do either of the following to eliminate the problem:

Use the dmt interleaving-delay command to increase the interleaving delay to 16 ms.
Use the dmt bitrate command to lower the maximum bit rates (both upstream and downstream, if necessary) to a lower rate than the one at which the errors were observed. [CSCdm82076]

Check Bytes Are Not Reported Correctly

The value you configure for the check bytes parameter might not be properly reported by the system. An incorrect check bytes value equal to half of the configured value is reported when a port is set to use standard training mode and the downstream data rate is greater than 7 Mbps. Only the downstream path is affected. The actual ratio of user data to check bytes matches the configuration. [CSCdm90741]

Interleaving Delay Is Not Reported Correctly

When a line that is configured for quick training mode trains to a downstream bit rate of 8 Mbps or higher, the actual interleaving delay might be reported incorrectly. (The interleaving delay appears in the display for the command show dsl interface atm slot/port.) When this problem occurs, the reported interleaving delay is half of the value that is actually in effect. [CSCdp14712]

Bitswap Command Is Not Functional

The dmt bitswap margin command has no effect in this release. Bit swapping is always disabled. [CSCdm80069]

Virtual Path Deconfiguration Problem

Sometimes it is impossible to remove a configured PVP. When VP switching is configured from one subscriber port to another using the same VPI number on both the ports, the VP circuit is not set up correctly.

Workaround: Configure different VPI numbers on the subscriber ports at either end of the connection. [CSCdp44469]

ATM Input Queue is Limited to 255 Cells

The maximum size of input queue counter that can be supported by hardware is 255 cells. (You can configure this using the max-size argument of the atm input-queue command.) But software allows configuration of values beyond 255. Limiting input traffic will work only for queue sizes configured 255 cells or less. When you enter a value higher than 255, no input queue function is performed—all traffic is allowed through. [CSCdp25887]

Resolved Caveats—Release 12.1(7)DA3

The problems listed in Table 1 are fixed in Cisco IOS Release 12.1(7)DA3.

Table 1: Problems Resolved in Release 12.1(7)DA3

Caveat Number	Description
CSCdw65903	An error can occur with management protocol processing. Please use the following URL for further information: http://www.cisco.com/cgi-bin/bugtool/onebug.pl?bugid=CSCdw65903

Caveat Number

Description

CSCdw65903

An error can occur with management protocol processing. Please use the following URL for further information:

http://www.cisco.com/cgi-bin/bugtool/onebug.pl?bugid=CSCdw65903

Resolved Caveats—Release 12.1(7)DA2

The problems listed in Table 2 are fixed in Cisco IOS Release 12.1(7)DA2.

Table 2: Problems Resolved in Release 12.1(7)DA2

Caveat Number	Description
CSCdt11503	Unusual field format in SNMP results in unexpected action.
CSCdu22277	Intermittently, one or more of the line cards loses communication to the NI-2.
CSCdu25651	Redundant NI-2 never reaches the standby state.
CSCdu39185	Cisco 677 CPEs do not train at low rates with 8xDMTs.
CSCdu45628	The system may crash as a result of resetting the time of day clock on the system. Under most circumstances, the system will not crash.

Resolved Caveats—Release 12.1(7)DA

The problems listed in Table 3 are fixed in Cisco IOS Release 12.1(7)DA.

Table 3: Problems Resolved in Release 12.1(7)DA

Caveat Number	Description
CSCdr22795	During normal operation of the NI2, the system causes runt packet messages to be logged.
CSCdr34573	The CLI service option is inconsistently displayed for FR PVCs & Soft-VCs.
CSCdr39074	The Frame Relay connection traffic table (FR-CTT) parameters, CIR & PIR values of the default row (100) are modified to match the default line rate (128kbps) of IDSL cards.
CSCdr76996	The fan traps from the Cisco 6130 and Cisco 6260 are not of the same severity.
CSCds54602	With the 4xflexi-DMT configured, the DSLAM reports the upstream margin of -63 dB or 63 dB (depending on the DSP firmware version) when operating with Alcatel-based CPEs (tested with Efficient 5260 and 5262).
CSCds62201	High latency is noticed for a reply from the system to a SET operation when provisioning a 4x flexi from CAP to DMT and vice versa. This results in the SNMP utility timing out on the SET operation.
CSCds64436	SNMP walk loops in csfRecordTable (CISCO-ATM-SIG-DIAG-MIB) conditions.
CSCds68212	The ATM filter created from CLI does not show up in the MIB filter tables. An ATM filter cannot be created from MIB.
CSCds75700	Performance and interval counters for sdsl LOSQ are incorrect; however, cSdslStucPerfLOSQs is correct.
CSCdt02293	The value of adslAtucPerfInits is sometimes incorrect for DMT line cards. The error has only been detected for untrained ports so far.
CSCdt02507	ADSL DMT line cards sometimes return a large value (for example, 65535) for LPRs, LOFs, and LOSs. The total counter adslAturPerf LPRs, LOFs, and LOSs does not show this irregularity.
CSCdt03362	Setting ceAssetFru returns 0 for false; the correct value for false is 2.
CSCdt08345	The MIB walk starts to loop at interface 6. Only IDSL cards are affected.
CSCdt16929	Currently, the Cisco 677 does not support trellis encoding. When trellis encoding is enabled, in conjunction with the Cisco 677 CPE, performance may be impacted negatively.
CSCdt35500	EIGRP does not work properly on the redundant topology of networks with DSLAMs running code from Cisco IOS 12.1DA. It gives a nonoptimal path the same distance as an optimal path and puts incorrect routes into the routing table.
CSCdt37180	E1 entries are missing from the IF table. Getbulk request is not returning all rows when some columns are deprecated.
CSCdt45154	Excessive upstream physical layer errors are generated, in both directions, when 4-port DMT Rev 2 line cards are used with ITEX CPE at very short loop lengths (<500m).
CSCdt73914	Link up and link down traps are enabled by default.
CSCdt74114	Link up/link down traps must be enabled/disabled on a per interface basis.
CSCdt79972	When attempting to set the maximum or minimum fast path upstream or downstream data rate MIB objects (adslAtucChanConfFastMinTxRate and adslAtucChanConfFastMaxTxRate) in the ADSL line MIB, returned results will indicate that the MIB objects are not available. The maximum and minimum fast path upstream and downstream data rates are configurable via profile settings in the NI-2 CLI.
CSCdt82888	Threshold temperature used to control fan speed is low for Cisco 6160 new fan.
CSCdt87067	When using the 8xDMT card in fast mode on loops up to 8 kft, the achievable downstream data rate may be observed to vary from port to port.
CSCdu03437	The minimum bit rate setting will be lost after reload.
CSCdu05353	An SNMP query of DSLAM with cSdslLineConfProfile causes an unexpected system reload.
CSCdu07097	When you initiate a show running configuration command from the NI-2, the active LED on some 8xDMT and G.SHDLS cards may be observed to turn off briefly. Active ports on these cards may also experience data loss during this period.

Resolved Caveats—Release 12.1(6)DA2

The problems listed in Table 4 are fixed in Cisco IOS Release 12.1(6)DA2

Table 4: Problems Resolved in Release 12.1(6)DA2

Caveat Number	Description
CSCdu07097	When you initiate a show running configuration command from the NI-2, the active LED on some 8xDMTs may be observed to turn off briefly. Active ports on these cards may also experience data loss during this period.

Resolved Caveats—Release 12.1(6)DA1

The problems listed in Table 5 are fixed in Cisco IOS Release 12.1(6)DA1.

Table 5: Problems Resolved in Release 12.1(6)DA1

Caveat Number	Description
CSCds80783	With 4xflexi-DMT and Cisco 677 CPE actual interleave delay differs from configured interleave delay.
CSCdt81792	The bit rate/reach performance on the 4xflexi-DMT is degraded.

Resolved Caveats—Release 12.1(6)DA

The problems listed in Table 6 are fixed in Cisco IOS Release 12.1(6)DA.

Table 6: Problems Resolved in Release 12.1(6)DA

Caveat Number	Description
CSCdm90741	Check bytes are not reported correctly.
CSCdr06985	VCs are not released when a flexi card changes type.
CSCdr36372	If you specify the UPC pass optional parameter in the frame-relay soft-vc... command, the parameter appears in the running configuration, even though this is the default.
CSCdr62008	The default value for the optional efci-bit parameter of the IDSL interface configuration command frame-relay soft-vc... is incorrect. The default should be 0, not map-fecn.
CSCdr92983	Occasionally, an SNMP query of the cIdslIturPerfPrev1DayMoniSecs object in the CISCO-IDSL-LINE-MIB results in a negative value.
CSCds06327	Under certain conditions, if you issue the no frame-relay soft-vc dlci, another frame-relay soft-vc on that interface will be improperly be deleted. This behavior only applies to IDSL interfaces.
CSCds26889	Two NI-2s have the same ATM address.
CSCds59103	ATM performance data for SONET/DS3 interfaces is reported in the physical layer.
CSCds61302	Setting ceAlarmHistTableSize returns ERROR and reboots the system.
CSCds74434	The VCCrossConnectTable in the ATM-MIB does not list all entries after a slot is deleted.
CSCds77253	The F4 and F5 OAM ATM cells arriving on modem interfaces are ignored.
CSCds80219	Entries for the trunk and subtend ports in the pnniIfTable are mapped to the wrong layer.
CSCds83050	Cisco IOS fails to report transmission rate correctly when line is down.
CSCds83174	If the interleaving delay for a CAP port is set to short, the configuration will not survive if the system is rebooted.
CSCds86503	The show facility station alarm shows LOS, LOF, or LOCD failure on a flexi-CAP line card whose interface has been shut down.
CSCdt15896	The values for the ciscoAtmIfXmitLed and ciscoAtmIfRecvLed in the CISCO-ATM-IF-MIB are incorrect for DSL ports.
CSCdt25661	There is an inconsistency between CLI and MIB in Cisco IOS on ILMI status.
CSCdt25829	Configuring an in-band management PVC (such as atm pvc 0/32) on the CPU interface (such as interface atm0/0.332) may cause the system to reload with an arithmetic exception.
CSCdt32735	OIR causes PVC/PVP duplication for line card interfaces.
CSCdt32735	The PVCs and PVPs created on a line card ATM interface would be duplicated if that slot has been reconfigured after deleting it and configuring a new slot.

Resolved Caveats—Release 12.1(5)DA

The problems listed in Table 7 are fixed in Cisco IOS Release 12.1(5)DA.

Table 7: Problems Resolved in Release 12.1(5)DA

Caveat Number	Description
CSCdp03232	No error message is generated when you try to delete a DMT profile that is in use.
CSCdp21731	The system incorrectly allows you to create VBR-RT and VBR-NRT tunnels. In fact the system can handle only CBR tunnels, and tunnels created as VBR-RT or VBR-NRT behave like CBR tunnels.
CSCdp68075	The input queue field in the output of the show interface atm slot#/port# command is not applicable in the NI-2 environment.
CSCdr01563	The ping atm oam command does not parse correctly in exec mode.
CSCdr18239	When autoconfiguration is enabled on a subscriber port, the DSLAM might set the maximum VCI bits for that port to an incorrect value that exceeds the maximum allowed value of 8. This depends upon a negotiated value with the connecting port on the adjacent node.
CSCdr19173	The software check for invalid VPI numbers on VP tunnels occurs only if the interface associated with the VP tunnel is up. If the interface is down, the system accepts invalid VPI values, but when the interfaces come back up, VPIs are assigned unpredictably.
CSCdr27204	Ping command fails in extended command line.
CSCdr32135	SNMP agent does not perform correct lexicographic ordering with instance-IDs padded with excess subidentifiers, while querying the cSdslLineConfProfileTable.
CSCdr32135	SNMP agent does not perform correct lexicographic ordering with instance-IDs padded with excess sub-identifiers, while querying the cSdslLineConfProfileTable.
CSCdr45065	SNMP queries to MIB variables in cSslStucIntervalTable and cSdslStucPerfDataTable (CISCO-SDSL-LINE-MIB) return incorrect values.
CSCdr63938	There is a performance issue on Alcatel-Based CPEs when the line length is less than 8 K feet with LOW noise, and the downstream bit rate is less than 512 kbps.
CSCdr89374	Reload fails with uncompression error.
CSCds33094	Error when using HP OV loading the mib.
CSCds36366	The DSLAM System reloads when acctngInterfaceEnable in the accountingControlMIB is set to true.
CSCds36366	DSLAM reloads when acctingInterfaceEnable is set to true.
CSCds36465	The atm oam intercept end-to-end command is not supported.
CSCds47433	Setting new values for multiple DS3/E3 attributes returns success, but the attributes retain old values.
CScds49329	SNMP queries to validInterval/invalidIntervals MIB variables in CISCO-SDSL-LINE-MIB return incorrect values.
CSCds49372	The pnniMapNode objects return incorrect values.
CSCds49743	Some entries in the ceAlarmTable, do not correspond to entries in the entPhysicalTable.
CSCds62190	VCLs do not show up in the atmVclTable if vci = 0.
CSCds71557	DMT line card performance data is not accessible.

Resolved Caveats—Release 12.1(4)DA

The problems listed in Table 8 are fixed in Cisco IOS Release 12.1(4)DA.

Table 8: Problems Resolved in Release 12.1(4)DA

Caveat Number	Description
CSCdp61780	No matter what you set the configuration register or the bootvar to, rommon always boots up off bootflash.
CSCdr58384	The line card operational mode reports G.DMT when the operational mode is actually T1.413.
CSCdr64678	If you disable snmp-server traps alarms with the no snmp-server enable traps alarms command, you cannot reenable this feature with the snmp-server enable traps alarms command.
CSCdr66665	After you upgrade a DSLAM from Cisco IOS Release 12.0(x)DA to 12.1(x)DA, you might experience a problem that results from a change in some of the DMT defaults in the DSL profiles. The changes affect the training mode, overhead framing, and DMT margin parameters.
CSCdr91551	The ITUC-1-8IDSL line card is not forwarding traffic after reload/reboot because the ISDN chipset is initializing indefinitely due to network clock synchronization.
CSCdr91706	Router hangs if the Cisco IOS HTTP service is enabled and you browse to an invalid URL.
CSCdr94959	NI-2/6160 allows the CPE to train with an incorrect mode.
CSCds15858	This unexpected behavior affects users who have a line card populated in slot 32. The soft VC addresses of ports atm32/1, atm32/2, and atm32/3 are identical to the ATM trunk ports atm0/1, atm0/2 and atm0/3 (in the case of an NI-2 with 2 subtend ports). This causes problems with the configuration information reporting in the wrong location, as well as errors with PNNI and possibly other protocols. The symptom occurs on the Cisco 6100, Cisco6130, Cisco 6160 and Cisco6260 running Cisco IOS Release 12.1(3)DA.

Resolved Caveats—Release 12.1(3)DA

The problems listed in Table 9 are fixed in Cisco IOS Release 12.1(3)DA.

Table 9: Problems Resolved in Release 12.1(3)DA

Caveat Number	Description
CSCdm88409	If you momentarily power down and then power up a CPE that is attached to a 4xDMT over ISDN, there is a small chance (roughly 1 in 100) that it will take an additional 45 seconds to complete training (as compared to typical training cycles).
CSCdm64615	If the 4xDMT over ISDN boot register is set for the card to reload application code, it reloads when the system reboots even if the load image is the same as the code already loaded on the card. (The reload takes about 2 minutes.)
CSCdr66442	A device that is connected to an 8xIDSL interface and that requires a unified network clock might experience CRC errors, because the 8xIDSL interface currently uses a local clock only. Support for using a network clock is planned for a future release.
CSCdr04963	The8xIDSL does not achieve theoretically possible downstream throughput for small frames (frames on the order of 64 bytes). The line card does achieve theoretically possible downstream rates for larger frames (frames on the order of 256 bytes and up).
CSCdr67353	The output of the show controllers command for a port that is configured for PPP encapsulation does not show the VPI/VCI that is configured between the NI-2 and the line card for that interface. Instead, it always shows VPI/VCI = 0/0, with all the SAR counters being 0. (The count for number of VCs configured is correct.)

Resolved Caveats—Release 12.1(2)DA

The problems listed in Table 10 are fixed in Cisco IOS Release 12.1(2)DA.

Table 10: Problems Resolved in Release 12.1(2)DA

Caveat Number	Description
CSCdp05249	When a hardware failure occurs on one or more of the ports on a 4xDMT, every other (otherwise healthy) port on the line card might be unable to train. Cisco strongly recommends that you replace or remove any line card with any defective ports.
CSCdr56410	When IP routing is enabled on an NI-2 DSLAM, the Ethernet port 0/0 does not come back up by itself after it experiences a problem.
CSCdm89206	Default ATM traffic descriptors cannot be deleted. When you attempt to delete them through the MIB, the MIB incorrectly indicates that the deletion is successful.

Resolved Caveats—Release 12.1(1)DA1

The problems listed in Table 11 are fixed in Cisco IOS Release 12.1(1)DA1.

Table 11: Problems Resolved in Release 12.1(1)DA1

Caveat Number	Description
CSCdr43418	Frames are delayed coming downstream from the NI-2 card.
CSCdr53955	ILMI does not come up between two subtended nodes. As a result, some traffic is dropped in error.

Resolved Caveats—Release 12.1(1)DA

The problems listed in Table 12 are fixed in Cisco IOS Release 12.1(1)DA.

Table 12: Problems Resolved in Release 12.1(1)DA

Caveat Number	Description
CSCdp21747, CSCdp03335	The CLI command for creating soft VPs, atm soft-vp, shows an invalid VPI range, and the MIB allows invalid values to be set. The full range of VPI values, 0 to 255, is now supported.
CSCdp19413	The dmtCodewordSize object does not allow a setting of automatic, as the CLI does.
CSCdm92046	The clear counter command does not consistently return all counter values to zero. This problem is not reproducible.
CSCdp17666	When you execute the shutdown command for a port, the port goes out of service, but Cisco IOS still lists the port's administrative status as UP.
CSCdp00996	Periodically, the system generates a serial management bus cyclic redundancy check (SMBCRC) error message such as this: `%NI2-5-LC_NOTICE:Slot[15] SMBCRC (1 days, 3 hours, 37 minutes, 6 seconds)` These messages are informational only and do not affect performance.
CSCdm66087	DMT lines do not train to upstream bit rates that are multiples of 128 kbps. This applies to both standard training and quick training modes, but only in the upstream direction. If a line is configured with a maximum upstream bit rate that is a multiple of 128 kbps, the best upstream bit rate to which the line can train is 32 kbps less than the configured value.
CSCdm91877	When a port is set to use quick training mode, its symbols per codeword value is sometimes reported as "?" (a question mark). This value is displayed when you enter the show dsl interface atm command.
CSCdp18533	The MIB defines the ifAlias (subscriber ID) object as a 64-character field. SNMP truncates the field at 20 characters. This problem is not reproducible.
CSCdp26405	If the BITS plug is configured as the system's timing source, and the BITS clock source fails, an alarm is generated. (This behavior is correct.) However, the alarm is not cleared even after you remove the BITS plug from the valid list of clock sources.
CSCdp63686	The system does not allow you to remove VC/VP configuration parameters from the table, falsely claiming the parameters are in use by a connection. The command that fails is no atm connection-traffic-table-row. This problem is not reproducible.

Obtaining Documentation

The following sections provide sources for obtaining documentation from Cisco Systems.

World Wide Web

You can access the most current Cisco documentation on the World Wide Web at the following sites:

http://www.cisco.com
http://www-china.cisco.com
http://www-europe.cisco.com

Documentation CD-ROM

Cisco documentation and additional literature are available in a CD-ROM package that ships with your product. The Documentation CD-ROM is updated monthly and is available as a single unit or through an annual subscription.

Registered Cisco.com users can order the Documentation CD-ROM through the online Subscription Store:

: http://www.cisco.com/go/subscription

Nonregistered Cisco.com users can order the Documentation CD-ROM through a local account representative by calling Cisco corporate headquarters (California, USA) at 408 526-7208 or, in North America, by calling 800 553-NETS(6387).
For assistance with ordering, contact your distributor or see the Cisco Technical Assistance Center website at:

: http://www.cisco.com/tac

Documentation Feedback

If you are a registered Cisco.com user, you can submit technical comments electronically.

To report problems, click Feedback in the toolbar at the top of any document web page and select Report a Problem. Under Problem Type, select CD/Online Documentation. After you complete the form, click Submit to send it to Cisco.

: You can also e-mail your comments to bug-doc@cisco.com .

To send your opinion of the DSL documentation, go to the following URL:

: http://www.cisco.com/univercd/cc/td/doc/product/dsl_prod/
: and click here at the top of the page in the sentence "Is this Cisco documentation helpful? Click here to give us your feedback." This produces a Documentation Survey that you can complete and send to Cisco electronically.

To submit your comments by mail, write to the following address:

Cisco Systems
Attn: Document Resource Connection
170 West Tasman Drive
San Jose, CA 95134-9883

We appreciate your comments.

Obtaining Technical Assistance

Cisco provides Cisco.com as a starting point for all technical assistance. Customers and partners can obtain documentation, troubleshooting tips, and sample configurations from online tools. For Cisco.com registered users, additional troubleshooting tools are available from the TAC website.

Cisco.com

Cisco.com is the foundation of a suite of interactive, networked services that provides immediate, open access to Cisco information and resources at anytime, from anywhere in the world. This highly integrated Internet application is a powerful, easy-to-use tool for doing business with Cisco.

Cisco.com provides a broad range of features and services to help customers and partners streamline business processes and improve productivity. Through Cisco.com, you can find information about Cisco and our networking solutions, services, and programs. In addition, you can resolve technical issues with online technical support, download and test software packages, and order Cisco learning materials and merchandise. Valuable online skill assessment, training, and certification programs are also available.

Customers and partners can self-register on Cisco.com to obtain additional personalized information and services. Registered users can order products, check on the status of an order, access technical support, and view benefits specific to their relationships with Cisco.

To access Cisco.com, go to the following website:

http://www.cisco.com

Technical Assistance Center

The Cisco TAC website is available to all customers who need technical assistance with a Cisco product or technology that is under warranty or covered by a maintenance contract.

Contacting TAC by Using the Cisco TAC Website

If you have a priority level 3 (P3) or priority level 4 (P4) problem, contact TAC by going to the TAC website:

http://www.cisco.com/tac

P3 and P4 level problems are defined as follows:

P3—Your network performance is degraded. Network functionality is noticeably impaired, but most business operations continue.
P4—You need information or assistance on Cisco product capabilities, product installation, or basic product configuration.

In each of the above cases, use the Cisco TAC website to quickly find answers to your questions.

To register for Cisco.com, go to the following website:

http://www.cisco.com/register/

If you cannot resolve your technical issue by using the TAC online resources, Cisco.com registered users can open a case online by using the TAC Case Open tool at the following website:

http://www.cisco.com/tac/caseopen

Contacting TAC by Telephone

If you have a priority level 1(P1) or priority level 2 (P2) problem, contact TAC by telephone and immediately open a case. To obtain a directory of toll-free numbers for your country, go to the following website:

http://www.cisco.com/warp/public/687/Directory/DirTAC.shtml

P1 and P2 level problems are defined as follows:

P1—Your production network is down, causing a critical impact to business operations if service is not restored quickly. No workaround is available.
P2—Your production network is severely degraded, affecting significant aspects of your business operations. No workaround is available.

Table of Contents

Release Notes for Cisco IOS Release 12.1(7)DA3 for Cisco DSLAMs with NI-2

Requirements

Limitations

Number of Configurable MPLS VPNs Limited to 50

Integrated Routing and Bridging Not Supported

VPN Interfaces Restricted to Trunk Interfaces

MPLS ATM-Label Switch Router Functionality Not Supported

Performance Restrictions for MPLS VPN Traffic

Restricted Layer 3 Services

Restricted MPLS Features

DSL Interface Limitations

Configuration of MPLS VPN Mapping of RFC 1483 Routed Sessions not Supported by Cisco DSL Manager (CDM)

MPLS VPN Mapping of RFC 1483 Routed Sessions not Supported on the Eight-Port IDSL ITU-C Line Card

Unexpected Reload When Querying cIdslPppoa tables

Autosync of Running Config Can Cause Write Command to Fail

Some DSL Statistics Missing in 12.1(5)DA to 12.1(7)DA for DMT Cards

Minimum DMT Bit Rate Setting Lost After Upgrade

Profile Created from Altered Default has Diff Parameters

E1 IMA Group Continuously Loses Cells When Physical Link Removed

Upstream Errors on Some Alcatel CPE in Fast Mode

In Very Rare Cases after Power Cycle a Port Will not Pass Data

Far End Statistics Reported as Zero

In Extremely Rare Cases After a Power Cycle, Port does not Train

Attenuation Reporting Zero Through CLI and SNMP

1544 Annex B Line Rate Does not Meet PSD Mask Limit

The 8xG.SHDSL Line Card does not Support Rate Adaption

Active LED on 8xDMT & 8xG.SHDSL Stays on When NI-2 Communication Lost

IDSL Traffic Stops when a Second NI-2 Card is Inserted in a Chassis

Link Down Traps not Reliably Sent for IDSL

No Log Event Link-status has no Effect

Link up Traps Still Mapped to ATM Interface

The Delay does not Change when Changing oamLoopbackPingDelay.

adslAturChanTransmittedBlks has a MIB value of 0 while CLI has non-0

adslAturChanReceivedBlks has a MIB value of 0 and CLI has non-0

adslAturVendorID will return 0 from MIB, CLI has 57 (Globespan)

adslAtucChanCrcBlockLength always has a value of 0

adslAturChanCrcBlockLength always has a value of 0

cIdslItucInvSerialNumber returns garbage instead of empty string

On OC-3 NI-2s, PPP Sessions May Drop During Switchovers

Peer NI2 Will Reboot When Fuse Blows on Other NI-2

Redundant NI-2s Reload When All DSL Ports Drop Simultaneously

Autosync of Start-up Config Does Not Always Work

677 3.0 Cat1 Down Trains Lower Without Noise Than With

G.Lite Loss of Frames, CPE Will not Train

The 8xDMT Line Card Ignores Interleave Setting When Trained to 677

Profile Lists Rates of 1552, 2064, 2320 & Allows Their Use

677 2.x Reports Incorrect Up Margin for Cat1 Loop midCSA6

c677 w/3.0 FW and Cat1 Loop Anomaly with Injected Noise

Alcatel-Based CPEs Training Odd at High Bit Rates, Interleaved

Degraded Downstream Bit Rates at Different Lengths from 12.1(5)DA

4xFlexi-DMT Unable to Train to Maximum. Rates with 677 CPEs

Next: Low Reported Margin for 677 Firmware 2.2

Running-config Incorrectly Marked as Changed

ST Pro in G.lite Mode Fails Several Annex D Test Loops

Takes More Time Than Downshift Interval to Meet Rate Adaptation

Westell HEC Errors, Frame Loss for Low Margins (0 to 3)

Creation of snmpNotifyFilterProfileTable Incorrect—Can Cause Reboot

Filesys space Error When 1st Write Mem After Reload

677 2.x Reports Incorrect Up Margin for Test Loop midCSA6

8xDMT to 677 3.0 CSF, USF, and Margin Perf May Differ from Flexi

Actual Margin < Provisioned for G.lite and DMT

Alcatel Modems: no train when ds<=1024 and Trellis=disabled

Downstream Margin not Reported with Cisco 677 CPE (ADI ver. 2.2)

Downstream USF in Fast Path with Cisco 678

Fast Latency Will not Train Downstream < 64 kbps

Increased Corrected and Uncorrected SF Counts US with 5262 CPE

Increased Corrected and Uncorrected SF Counts with Westell CPE

LOF, LOS, LPR may be Reported Inaccurately

The Margin Reported with 678 CPE may Differ from Actual

Unable to Obtain Requested Int. Delay with /677

Upstream Margin not Accurate for Alcatel-based CPE and Cisco 678

Variable Train Rates with 677 CPE (12.5 to 13.0 kft)

CIB Save FAILED Message Displayed During NI Boot

DSLAM Lockup with Multiple Slot/no Slot Commands

FastPath Working on G.lite

BERT Test Interrupts Upstream Traffic on the Other Card Ports

Downstream Frame Loss When Check-bytes Less Than 6 with 5262/Interleave

adslAturCurrOutputPwr Returns Incorrect Value on CAP Cards

adslAtucCurrStatus Returns Incorrect Value for CAP and DMT